WO2023059136A1 - 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 - Google Patents
포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 Download PDFInfo
- Publication number
- WO2023059136A1 WO2023059136A1 PCT/KR2022/015149 KR2022015149W WO2023059136A1 WO 2023059136 A1 WO2023059136 A1 WO 2023059136A1 KR 2022015149 W KR2022015149 W KR 2022015149W WO 2023059136 A1 WO2023059136 A1 WO 2023059136A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- point cloud
- cloud data
- subgroup
- bitstream
- geometry
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/70—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by syntax aspects related to video coding, e.g. related to compression standards
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
- G06T9/001—Model-based coding, e.g. wire frame
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/119—Adaptive subdivision aspects, e.g. subdivision of a picture into rectangular or non-rectangular coding blocks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/129—Scanning of coding units, e.g. zig-zag scan of transform coefficients or flexible macroblock ordering [FMO]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/174—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a slice, e.g. a line of blocks or a group of blocks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/184—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being bits, e.g. of the compressed video stream
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/30—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using hierarchical techniques, e.g. scalability
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
- H04N19/537—Motion estimation other than block-based
- H04N19/54—Motion estimation other than block-based using feature points or meshes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/597—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding specially adapted for multi-view video sequence encoding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/90—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals
- H04N19/96—Tree coding, e.g. quad-tree coding
Definitions
- Embodiments relate to a method and apparatus for processing point cloud content.
- the point cloud content is content expressed as a point cloud, which is a set of points belonging to a coordinate system representing a 3D space.
- Point cloud content can express three-dimensional media, and provides various services such as VR (Virtual Reality), AR (Augmented Reality), MR (Mixed Reality), and autonomous driving service. used to provide However, tens of thousands to hundreds of thousands of point data are required to express point cloud content. Therefore, a method for efficiently processing a vast amount of point data is required.
- Embodiments provide an apparatus and method for efficiently processing point cloud data.
- Embodiments provide a point cloud data processing method and apparatus for solving latency and encoding/decoding complexity.
- a point cloud data transmission method includes encoding point cloud data; and transmitting a bitstream including point cloud data; can include
- a method for receiving point cloud data according to embodiments includes receiving a bitstream including point cloud data; and decoding the point cloud data; can include
- Apparatus and method according to embodiments may process point cloud data with high efficiency.
- Devices and methods according to embodiments may provide a point cloud service of high quality.
- Devices and methods according to embodiments may provide point cloud content for providing general-purpose services such as VR services and autonomous driving services.
- FIG. 1 shows an example of a point cloud content providing system according to embodiments.
- FIG. 2 is a block diagram illustrating an operation of providing point cloud content according to embodiments.
- FIG 3 shows an example of a point cloud video capture process according to embodiments.
- FIG. 4 shows an example of a point cloud encoder according to embodiments.
- FIG. 5 illustrates an example of a voxel according to embodiments.
- FIG. 6 shows an example of an octree and an occupancy code according to embodiments.
- FIG. 7 shows an example of a neighbor node pattern according to embodiments.
- FIG. 10 shows an example of a point cloud decoder according to embodiments.
- FIG. 11 shows an example of a point cloud decoder according to embodiments.
- FIG. 13 is an example of a receiving device according to embodiments.
- FIG. 14 shows an example of a structure capable of interworking with a method/apparatus for transmitting and receiving point cloud data according to embodiments.
- 15 illustrates a process of encoding, transmitting, and decoding point cloud data according to embodiments.
- FIG. 16 illustrates a layer-based point cloud data configuration according to embodiments and a geometry and attribute bitstream structure according to embodiments.
- 17 shows a bitstream configuration according to embodiments.
- 19 illustrates a method of selecting geometry data and attribute data according to embodiments.
- FIG. 20 illustrates a method of constructing a slice including point cloud data according to embodiments.
- FIG. 21 illustrates single slice and segmented slice-based geometry tree structures according to embodiments.
- FIG. 22 illustrates a layer group structure of a geometry coding tree and an aligned layer group structure of an attribute coding tree according to embodiments.
- FIG. 23 illustrates a layer group and subgroup structure according to embodiments.
- FIG. 27 illustrates a slice generation method considering cloud density according to embodiments.
- 29 illustrates a subgroup division process according to embodiments.
- FIG. 30 shows a comparison between a variable subgroup size and a fixed subgroup size according to embodiments.
- 31 illustrates a subgroup bounding box according to a node scan order according to embodiments.
- 32 illustrates a subgroup bounding box according to a node scan order according to embodiments.
- 33 shows a bitstream including point cloud data and parameters according to embodiments.
- 35 shows a dependent geometry data unit header according to embodiments.
- FIG. 36 shows a layer group structure inventory according to embodiments.
- FIG. 37 shows a point cloud data transmission device according to embodiments.
- 38 shows an apparatus for receiving point cloud data according to embodiments.
- 39 illustrates a sub-bitstream classification method according to embodiments.
- 40 illustrates a method for transmitting and receiving point cloud data according to embodiments.
- 41 illustrates a method for transmitting and receiving point cloud data according to embodiments.
- FIG. 43 illustrates a point cloud data transmission method according to embodiments.
- 44 shows a method for receiving point cloud data according to embodiments.
- FIG. 1 shows an example of a point cloud content providing system according to embodiments.
- the point cloud content providing system shown in FIG. 1 may include a transmission device 10000 and a reception device 10004.
- the transmitting device 10000 and the receiving device 10004 may perform wired/wireless communication to transmit/receive point cloud data.
- the transmission device 10000 may secure, process, and transmit point cloud video (or point cloud content).
- the transmission device 10000 may include a fixed station, a base transceiver system (BTS), a network, an artificial intelligence (AI) device and/or system, a robot, an AR/VR/XR device and/or a server. etc. may be included.
- BTS base transceiver system
- AI artificial intelligence
- the transmission device 10000 is a device that communicates with a base station and/or other wireless devices using a radio access technology (eg, 5G New RAT (NR), Long Term Evolution (LTE)), It may include robots, vehicles, AR/VR/XR devices, mobile devices, home appliances, Internet of Thing (IoT) devices, AI devices/servers, and the like.
- a radio access technology eg, 5G New RAT (NR), Long Term Evolution (LTE)
- NR 5G New RAT
- LTE Long Term Evolution
- It may include robots, vehicles, AR/VR/XR devices, mobile devices, home appliances, Internet of Thing (IoT) devices, AI devices/servers, and the like.
- IoT Internet of Thing
- the transmission device 10000 includes a point cloud video acquisition unit (Point Cloud Video Acquisition, 10001), a point cloud video encoder (Point Cloud Video Encoder, 10002), and/or a transmitter (or Communication module), 10003 ) include
- the point cloud video acquisition unit 10001 acquires a point cloud video through processing such as capture, synthesis, or creation.
- Point cloud video is point cloud content expressed as a point cloud, which is a set of points located in a 3D space, and may be referred to as point cloud video data.
- a point cloud video according to embodiments may include one or more frames. One frame represents a still image/picture. Accordingly, the point cloud video may include a point cloud image/frame/picture, and may be referred to as any one of a point cloud image, a frame, and a picture.
- the point cloud video encoder 10002 encodes secured point cloud video data.
- the point cloud video encoder 10002 may encode point cloud video data based on point cloud compression coding.
- Point cloud compression coding may include geometry-based point cloud compression (G-PCC) coding and/or video based point cloud compression (V-PCC) coding or next-generation coding.
- G-PCC geometry-based point cloud compression
- V-PCC video based point cloud compression
- point cloud compression coding is not limited to the above-described embodiments.
- the point cloud video encoder 10002 can output a bitstream containing encoded point cloud video data.
- the bitstream may include not only encoded point cloud video data, but also signaling information related to encoding of the point cloud video data.
- Transmitter 10003 transmits a bitstream containing encoded point cloud video data.
- a bitstream according to embodiments is encapsulated into a file or a segment (eg, a streaming segment) and transmitted through various networks such as a broadcasting network and/or a broadband network.
- the transmission device 10000 may include an encapsulation unit (or encapsulation module) that performs an encapsulation operation.
- the encapsulation unit may be included in the transmitter 10003.
- the file or segment may be transmitted to the receiving device 10004 through a network or stored in a digital storage medium (eg, USB, SD, CD, DVD, Blu-ray, HDD, SSD, etc.).
- the transmitter 10003 is capable of wired/wireless communication with the receiving device 10004 (or the receiver 10005) through a network such as 4G, 5G, or 6G.
- the transmitter 10003 may perform necessary data processing operations depending on the network system (for example, a communication network system such as 4G, 5G, or 6G).
- the transmission device 10000 may transmit encapsulated data according to an on demand method.
- a receiving device 10004 includes a receiver 10005, a point cloud video decoder 10006, and/or a renderer 10007.
- the receiving device 10004 is a device or robot that communicates with a base station and/or other wireless devices using a wireless access technology (eg, 5G New RAT (NR), Long Term Evolution (LTE)) , vehicles, AR/VR/XR devices, mobile devices, home appliances, Internet of Things (IoT) devices, AI devices/servers, and the like.
- a wireless access technology eg, 5G New RAT (NR), Long Term Evolution (LTE)
- the receiver 10005 receives a bitstream including point cloud video data or a file/segment in which the bitstream is encapsulated from a network or a storage medium.
- the receiver 10005 may perform necessary data processing operations depending on the network system (eg, 4G, 5G, 6G communication network system).
- the receiver 10005 may output a bitstream by decapsulating the received file/segment.
- the receiver 10005 may include a decapsulation unit (or decapsulation module) for performing a decapsulation operation.
- the decapsulation unit may be implemented as an element (or component) separate from the receiver 10005.
- the point cloud video decoder 10006 decodes a bitstream containing point cloud video data.
- the point cloud video decoder 10006 can decode the point cloud video data according to the way it was encoded (eg, the reverse of the operation of the point cloud video encoder 10002). Accordingly, the point cloud video decoder 10006 may decode point cloud video data by performing point cloud decompression coding, which is a reverse process of point cloud compression.
- Point cloud decompression coding includes G-PCC coding.
- a renderer 10007 renders the decoded point cloud video data.
- the renderer 10007 may output point cloud content by rendering audio data as well as point cloud video data.
- the renderer 10007 may include a display for displaying point cloud content.
- the display may not be included in the renderer 10007 and may be implemented as a separate device or component.
- the feedback information is information for reflecting the interactivity with the user consuming the point cloud content, and includes user information (eg, head orientation information), viewport information, etc.).
- user information eg, head orientation information
- viewport information etc.
- the feedback information is sent to the content transmitter (eg, the transmission device 10000) and/or the service provider.
- the feedback information may be used in the receiving device 10004 as well as in the transmitting device 10000, or may not be provided.
- Head orientation information is information about a user's head position, direction, angle, movement, and the like.
- the receiving device 10004 may calculate viewport information based on head orientation information.
- Viewport information is information about an area of a point cloud video that a user is looking at.
- a viewpoint is a point at which a user watches a point cloud video, and may mean a central point of a viewport area. That is, the viewport is an area centered on the viewpoint, and the size and shape of the area may be determined by FOV (Field Of View).
- FOV Field Of View
- the receiving device 10004 performs gaze analysis and the like to check the point cloud consumption method of the user, the point cloud video area that the user gazes at, the gaze time, and the like.
- the receiving device 10004 may transmit feedback information including the result of the gaze analysis to the transmitting device 10000.
- Feedback information according to embodiments may be obtained in a rendering and/or display process.
- Feedback information according to embodiments may be obtained by one or more sensors included in the receiving device 10004.
- feedback information may be secured by the renderer 10007 or a separate external element (or device, component, etc.).
- a dotted line in FIG. 1 indicates a process of transmitting feedback information secured by the renderer 10007.
- the point cloud content providing system may process (encode/decode) point cloud data based on the feedback information. Accordingly, the point cloud video data decoder 10006 may perform a decoding operation based on the feedback information. Also, the receiving device 10004 may transmit feedback information to the transmitting device 10000. The transmission device 10000 (or the point cloud video data encoder 10002) may perform an encoding operation based on the feedback information. Therefore, the point cloud content providing system does not process (encode/decode) all point cloud data, but efficiently processes necessary data (for example, point cloud data corresponding to the user's head position) based on feedback information, and Point cloud content can be provided to
- the transmitting apparatus 10000 may be referred to as an encoder, a transmitting device, a transmitter, and the like, and a receiving apparatus 10004 may be referred to as a decoder, a receiving device, and a receiver.
- Point cloud data (processed through a series of processes of acquisition/encoding/transmission/decoding/rendering) in the point cloud content providing system of FIG. 1 according to embodiments will be referred to as point cloud content data or point cloud video data.
- point cloud content data may be used as a concept including metadata or signaling information related to point cloud data.
- Elements of the point cloud content providing system shown in FIG. 1 may be implemented as hardware, software, processor, and/or a combination thereof.
- FIG. 2 is a block diagram illustrating an operation of providing point cloud content according to embodiments.
- the block diagram of FIG. 2 shows the operation of the point cloud content providing system described in FIG. 1 .
- the point cloud content providing system may process point cloud data based on point cloud compression coding (eg, G-PCC).
- point cloud compression coding eg, G-PCC
- the point cloud content providing system may obtain a point cloud video (20000).
- Point cloud video is expressed as a point cloud belonging to a coordinate system representing a three-dimensional space.
- a point cloud video according to embodiments may include a Ply (Polygon File format or the Stanford Triangle format) file. If the point cloud video has one or more frames, the acquired point cloud video may include one or more Ply files.
- Ply files include point cloud data such as geometry and/or attributes of points. Geometry contains positions of points.
- the position of each point may be expressed as parameters (eg, values of each of the X-axis, Y-axis, and Z-axis) representing a three-dimensional coordinate system (eg, a coordinate system composed of XYZ axes).
- Attributes include attributes of points (eg, texture information of each point, color (YCbCr or RGB), reflectance (r), transparency, etc.).
- a point has one or more attributes (or properties).
- a point may have one color attribute or two attributes, color and reflectance.
- geometry may be referred to as positions, geometry information, geometry data, and the like, and attributes may be referred to as attributes, attribute information, and attribute data.
- the point cloud content providing system obtains points from information (for example, depth information, color information, etc.) related to the acquisition process of the point cloud video. Cloud data is available.
- a point cloud content providing system may encode point cloud data (20001).
- the point cloud content providing system may encode point cloud data based on point cloud compression coding.
- point cloud data may include geometry and attributes of points.
- the point cloud content providing system may output a geometry bitstream by performing geometry encoding to encode geometry.
- the point cloud content providing system may output an attribute bitstream by performing attribute encoding for encoding attributes.
- a point cloud content providing system may perform attribute encoding based on geometry encoding.
- a geometry bitstream and an attribute bitstream according to embodiments may be multiplexed and output as one bitstream.
- a bitstream according to embodiments may further include signaling information related to geometry encoding and attribute encoding.
- a point cloud content providing system may transmit encoded point cloud data (20002).
- Point cloud data encoded as described in FIG. 1 may be expressed as a geometry bitstream and an attribute bitstream.
- the encoded point cloud data may be transmitted in the form of a bitstream together with signaling information related to encoding of the point cloud data (eg, signaling information related to geometry encoding and attribute encoding).
- the point cloud content providing system may encapsulate a bitstream transmitting encoded point cloud data and transmit the encoded point cloud data in the form of a file or segment.
- a point cloud content providing system may receive a bitstream including encoded point cloud data. Also, the point cloud content providing system (eg, the receiving device 10004 or the receiver 10005) may demultiplex the bitstream.
- the point cloud content providing system may decode encoded point cloud data (eg, a geometry bitstream, an attribute bitstream) transmitted as a bitstream. there is.
- the point cloud content providing system eg, the receiving device 10004 or the point cloud video decoder 10005
- the point cloud content providing system eg, the receiving device 10004 or the point cloud video decoder 10005
- the point cloud content providing system may restore attributes of points by decoding an attribute bitstream based on the restored geometry.
- the point cloud content providing system (eg, the receiving device 10004 or the point cloud video decoder 10005) may reconstruct the point cloud video based on the decoded attributes and positions according to the reconstructed geometry.
- a point cloud content providing system may render the decoded point cloud data (20004).
- the point cloud content providing system eg, the receiving device 10004 or the renderer 10007) may render the geometry and attributes decoded through the decoding process according to various rendering methods. Points of the point cloud content may be rendered as a vertex with a certain thickness, a cube with a specific minimum size centered at the vertex position, or a circle centered at the vertex position. All or part of the rendered point cloud content is provided to the user through a display (eg, VR/AR display, general display, etc.).
- a display eg, VR/AR display, general display, etc.
- the point cloud content providing system (eg, the receiving device 10004) according to embodiments may secure feedback information (20005).
- the point cloud content providing system may encode and/or decode point cloud data based on the feedback information. Since the feedback information and operation of the point cloud content providing system according to the embodiments are the same as the feedback information and operation described in FIG. 1, a detailed description thereof will be omitted.
- FIG 3 shows an example of a point cloud video capture process according to embodiments.
- FIG. 3 shows an example of a point cloud video capture process of the point cloud content providing system described in FIGS. 1 and 2 .
- Point cloud content is a point cloud video (images and/or videos) are included.
- a point cloud content providing system includes one or more cameras (eg, an infrared camera capable of securing depth information, color information corresponding to depth information) to generate point cloud content.
- Point cloud video can be captured using an RGB camera, etc.), a projector (eg, an infrared pattern projector to secure depth information), or LiDAR.
- a system for providing point cloud content according to embodiments may secure point cloud data by extracting a shape of a geometry composed of points in a 3D space from depth information and extracting an attribute of each point from color information.
- Images and/or videos according to embodiments may be captured based on at least one of an inward-facing method and an outward-facing method.
- the left side of FIG. 3 shows the inward-facing method.
- the inward-pacing method refers to a method in which one or more cameras (or camera sensors) located around the central object capture the central object.
- the inward-pacing method is a point cloud content that provides users with 360-degree images of key objects (e.g., provides users with 360-degree images of objects (e.g., key objects such as characters, players, objects, actors, etc.) It can be used to create VR / AR content).
- the right side of FIG. 3 shows the outward-facing method.
- the outward-pacing method refers to a method in which one or more cameras (or camera sensors) located around a central object capture an environment of the central object other than the central object.
- the outward-facing method may be used to generate point cloud content (eg, content representing an external environment that may be provided to a user of an autonomous vehicle) for providing a surrounding environment from a user's point of view.
- point cloud content eg, content representing an external environment that may be provided to a user of an autonomous vehicle
- point cloud content may be generated based on a capture operation of one or more cameras.
- the point cloud content providing system may perform calibration of one or more cameras to set a global coordinate system before a capture operation.
- the point cloud content providing system may generate point cloud content by synthesizing an image and/or video captured by the above-described capture method and an arbitrary image and/or video.
- the point cloud content providing system may not perform the capture operation described in FIG. 3 when generating point cloud content representing a virtual space.
- the point cloud content providing system may perform post-processing on captured images and/or videos. That is, the point cloud content providing system removes an unwanted area (for example, the background), recognizes a space where captured images and/or videos are connected, and fills in a spatial hole if there is one. can
- the point cloud content providing system may generate one point cloud content by performing coordinate system conversion on points of the point cloud video obtained from each camera.
- the point cloud content providing system may perform coordinate system conversion of points based on the positional coordinates of each camera. Accordingly, the point cloud content providing system may generate content representing one wide range or point cloud content having a high density of points.
- FIG. 4 shows an example of a point cloud encoder according to embodiments.
- FIG. 4 shows an example of the point cloud video encoder 10002 of FIG. 1 .
- the point cloud encoder converts point cloud data (eg, positions of points and/or attributes) and perform encoding operations.
- point cloud data eg, positions of points and/or attributes
- the point cloud content providing system may not be able to stream the corresponding content in real time. Therefore, the point cloud content providing system may reconstruct the point cloud content based on the maximum target bitrate in order to provide it according to the network environment.
- the point cloud encoder can perform geometry encoding and attribute encoding. Geometry encoding is performed before attribute encoding.
- the point cloud encoder includes a transformation coordinates (40000), a quantization unit (Quantize and Remove Points (Voxelize), 40001), an analyze octree (40002), and a surface approximate analysis unit. (Analyze Surface Approximation, 40003), Arithmetic Encode (40004), Reconstruct Geometry (40005), Transform Colors (40006), Transfer Attributes (40007), RAHT It includes a transform unit 40008, a LOD generator 40009, a lifting transform unit 40010, a quantize coefficients unit 40011, and/or an Arithmetic Encode 40012. .
- the coordinate system conversion unit 40000, the quantization unit 40001, the octree analysis unit 40002, the surface approximate analysis unit 40003, the Arithmetic encoder 40004, and the geometry reconstruction unit 40005 perform geometry encoding. can do.
- Geometry encoding according to embodiments may include octree geometry coding, direct coding, trisoup geometry encoding, and entropy encoding. Direct coding and trisup geometry encoding are applied selectively or in combination. Also, geometry encoding is not limited to the above example.
- a coordinate system conversion unit 40000 receives positions and converts them into a coordinate system.
- the positions may be converted into positional information in a 3D space (eg, a 3D space expressed in XYZ coordinates).
- Location information in a 3D space may be referred to as geometry information.
- a quantization unit 40001 quantizes geometry.
- the quantization unit 40001 may quantize points based on minimum position values of all points (for example, minimum values on each axis for the X axis, Y axis, and Z axis).
- the quantization unit 40001 multiplies the difference between the minimum position value and the position value of each point by a preset quatization scale value, and then performs a quantization operation to find the nearest integer value by performing rounding or rounding.
- one or more points may have the same quantized position (or position value).
- the quantization unit 40001 performs voxelization based on quantized positions to reconstruct quantized points.
- points of point cloud content may be included in one or more voxels.
- the quantization unit 40001 may match groups of points in the 3D space to voxels.
- one voxel may include only one point.
- one voxel may include one or more points.
- the position of the center of a corresponding voxel may be set based on the positions of one or more points included in one voxel. In this case, attributes of all positions included in one voxel may be combined and assigned to the corresponding voxel.
- the octree analyzer 40002 performs octree geometry coding (or octree coding) to represent voxels in an octree structure.
- the octree structure represents points matched to voxels based on an octal tree structure.
- the surface approximation analyzer 40003 may analyze and approximate an octree.
- Octree analysis and approximation is a process of analyzing to voxelize a region including a plurality of points in order to efficiently provide octree and voxelization.
- Arismetic encoder 40004 entropy encodes an octree and/or an approximated octree.
- the encoding method includes an Arithmetic encoding method.
- a geometry bitstream is created.
- Color conversion section 40006, attribute conversion section 40007, RAHT conversion section 40008, LOD generation section 40009, lifting conversion section 40010, coefficient quantization section 40011 and/or Arithmetic encoder 40012 performs attribute encoding.
- one point may have one or more attributes. Attribute encoding according to embodiments is equally applied to attributes of one point. However, when one attribute (for example, color) includes one or more elements, independent attribute encoding is applied to each element.
- Attribute encoding may include color transform coding, attribute transform coding, region adaptive hierarchical transform (RAHT) coding, interpolaration-based hierarchical nearest-neighbour prediction-prediction transform coding, and interpolation-based hierarchical nearest transform (RAHT) coding.
- RAHT region adaptive hierarchical transform
- RAHT interpolaration-based hierarchical nearest-neighbour prediction-prediction transform
- RAHT interpolation-based hierarchical nearest transform
- -neighbor prediction with an update/lifting step (Lifting Transform)) coding may be included.
- the above-described RAHT coding, predictive transform coding, and lifting transform coding may be selectively used, or a combination of one or more codings may be used.
- attribute encoding according to embodiments is not limited to the above-described example.
- the color conversion unit 40006 performs color conversion coding to convert color values (or textures) included in attributes.
- the color conversion unit 40006 may convert a format of color information (for example, convert RGB to YCbCr).
- An operation of the color conversion unit 40006 according to embodiments may be optionally applied according to color values included in attributes.
- the geometry reconstructor 40005 reconstructs (decompresses) an octree and/or an approximated octree.
- the geometry reconstructor 40005 reconstructs an octree/voxel based on a result of analyzing the distribution of points.
- the reconstructed octree/voxel may be referred to as reconstructed geometry (or reconstructed geometry).
- the attribute transformation unit 40007 performs attribute transformation to transform attributes based on positions for which geometry encoding has not been performed and/or reconstructed geometry. As described above, since attributes depend on geometry, the attribute conversion unit 40007 can transform attributes based on reconstructed geometry information. For example, the attribute conversion unit 40007 may transform an attribute of a point at a position based on a position value of a point included in a voxel. As described above, when the position of the central point of a voxel is set based on the positions of one or more points included in one voxel, the attribute conversion unit 40007 transforms attributes of one or more points. When tri-soup geometry encoding is performed, the attribute conversion unit 40007 may transform attributes based on tri-soup geometry encoding.
- the attribute conversion unit 40007 is an average value of attributes or attribute values (eg, color or reflectance of each point) of neighboring points within a specific position/radius from the position (or position value) of the center point of each voxel. Attribute conversion can be performed by calculating .
- the attribute conversion unit 40007 may apply a weight according to the distance from the central point to each point when calculating the average value. Therefore, each voxel has a position and a calculated attribute (or attribute value).
- the attribute conversion unit 40007 may search for neighboring points existing within a specific location/radius from the position of the center point of each voxel based on the K-D tree or the Morton code.
- the K-D tree is a binary search tree and supports a data structure that can manage points based on location so that a quick Nearest Neighbor Search (NNS) is possible.
- the Morton code is generated by expressing coordinate values (for example, (x, y, z)) representing the three-dimensional positions of all points as bit values and mixing the bits. For example, if the coordinate value indicating the position of the point is (5, 9, 1), the bit value of the coordinate value is (0101, 1001, 0001).
- the attribute conversion unit 40007 may sort points based on Molton code values and perform a nearest neighbor search (NNS) through a depth-first traversal process. After the attribute transformation operation, if a nearest neighbor search (NNS) is required in another transformation process for attribute coding, a K-D tree or Morton code is used.
- NSS nearest neighbor search
- the converted attributes are input to the RAHT conversion unit 40008 and/or the LOD generation unit 40009.
- the RAHT conversion unit 40008 performs RAHT coding for predicting attribute information based on reconstructed geometry information. For example, the RAHT converter 40008 may predict attribute information of a node at a higher level of the octree based on attribute information associated with a node at a lower level of the octree.
- An LOD generator 40009 generates a level of detail (LOD) to perform predictive transformation coding.
- LOD according to embodiments is a degree representing detail of point cloud content. A smaller LOD value indicates lower detail of point cloud content, and a larger LOD value indicates higher detail of point cloud content. Points can be classified according to LOD.
- the lifting transform unit 40010 performs lifting transform coding for transforming attributes of a point cloud based on weights. As described above, lifting transform coding may be selectively applied.
- the coefficient quantization unit 40011 quantizes attribute-coded attributes based on coefficients.
- the Arithmetic Encoder 40012 encodes the quantized attributes based on Arithmetic Coding.
- One or more processors may perform at least one or more of operations and/or functions of elements of the point cloud encoder of FIG. 4 described above. Also, one or more processors may operate or execute a set of software programs and/or instructions to perform operations and/or functions of elements of the point cloud encoder of FIG. 4 .
- One or more memories may include high speed random access memory, and may include non-volatile memory (eg, one or more magnetic disk storage devices, flash memory devices, or other non-volatile solid state memory devices). memory devices (Solid-state memory devices, etc.).
- FIG. 5 illustrates an example of a voxel according to embodiments.
- voxel 5 is an octree structure that recursively subdivides a cubical axis-aligned bounding box defined by two poles (0,0,0) and (2 d , 2 d , 2 d ). It shows an example of a voxel generated through One voxel includes at least one point. Spatial coordinates of a voxel may be estimated from a positional relationship with a voxel group. As described above, a voxel has an attribute (color or reflectance, etc.) like a pixel of a 2D image/video. Since a detailed description of the voxel is the same as that described in FIG. 4, it will be omitted.
- FIG. 6 shows an example of an octree and an occupancy code according to embodiments.
- the point cloud content providing system (point cloud video encoder 10002) or the point cloud encoder (eg, octree analyzer 40002) efficiently manages a voxel area and/or position. To do so, octree structure-based octree geometry coding (or octree coding) is performed.
- FIG. 6 shows an octree structure.
- a 3D space of point cloud content according to embodiments is represented by axes (eg, X-axis, Y-axis, and Z-axis) of a coordinate system.
- the octree structure is created by recursively subdividing a cubical axis-aligned bounding box defined by the two poles (0,0,0) and (2 d , 2 d , 2 d ). . 2d may be set to a value constituting the smallest bounding box enclosing all points of the point cloud content (or point cloud video).
- d represents the depth of the octree.
- the value of d is determined according to the following formula. In the following equation, (x int n , y int n , z int n ) represents positions (or position values) of quantized points.
- the entire 3D space can be divided into 8 spaces according to division.
- Each divided space is represented by a cube with six faces.
- each of the eight spaces is further divided based on the axes of the coordinate system (for example, the X-axis, Y-axis, and Z-axis). Therefore, each space is further divided into eight smaller spaces.
- the divided small space is also expressed as a cube with six faces. This division method is applied until the leaf node of the octree becomes a voxel.
- the lower part of Fig. 6 shows the occupancy code of the octree.
- the octree's occupancy code is generated to indicate whether each of eight divided spaces generated by dividing one space includes at least one point. Therefore, one occupancy code is represented by 8 child nodes. Each child node represents the occupancy of the divided space, and the child node has a value of 1 bit. Therefore, the occupancy code is expressed as an 8-bit code. That is, if at least one point is included in a space corresponding to a child node, the corresponding node has a value of 1. If a point is not included in the space corresponding to a child node (empty), the corresponding node has a value of 0. Since the occupancy code shown in FIG.
- a point cloud encoder (eg, the Arismetic encoder 40004) according to embodiments may entropy-encode an occupancy code. Also, to increase compression efficiency, the point cloud encoder may intra/inter code the occupancy code.
- a receiving device (eg, the receiving device 10004 or the point cloud video decoder 10006) according to embodiments reconstructs an octree based on an occupancy code.
- a point cloud encoder (eg, the point cloud encoder of FIG. 4 or the octree analyzer 40002) may perform voxelization and octree coding to store positions of points.
- points in the 3D space are not always evenly distributed, there may be a specific area where many points do not exist. Therefore, it is inefficient to perform voxelization on the entire 3D space. For example, if few points exist in a specific area, there is no need to perform voxelization to that area.
- the point cloud encoder does not perform voxelization on the above-described specific region (or nodes other than leaf nodes of the octree), but directly codes the positions of points included in the specific region. ) can be performed. Coordinates of direct coding points according to embodiments are called a direct coding mode (DCM). Also, the point cloud encoder according to embodiments may perform trisoup geometry encoding for reconstructing positions of points in a specific area (or node) based on a voxel based on a surface model. Tri-Sup geometry encoding is a geometry encoding that expresses the representation of an object as a series of triangle meshes.
- a point cloud decoder can generate a point cloud from a mesh surface.
- Direct coding and trisup geometry encoding according to embodiments may be selectively performed. Also, direct coding and triangle geometry encoding according to embodiments may be performed in combination with octree geometry coding (or octree coding).
- the option to use direct mode to apply direct coding must be activated.
- the node to which direct coding is applied is not a leaf node, points must exist.
- the number of all points subject to direct coding must not exceed a predetermined limit. If the above condition is satisfied, the point cloud encoder (or the Arithmetic encoder 40004) according to the embodiments may entropy code positions (or position values) of points.
- the point cloud encoder (for example, the surface approximate analysis unit 40003) according to the embodiments determines a specific level (when the level is smaller than the depth d of the octree) of the octree, and from that level uses the surface model to determine the node Tri-sup geometry encoding for reconstructing the position of a point in an area based on voxels may be performed (tri-sup mode).
- the point cloud encoder may designate a level to which tri-sup geometry encoding is applied. For example, if the specified level is equal to the depth of the octree, the point cloud encoder does not operate in tri-sup mode.
- the point cloud encoder may operate in tri-sup mode only when the designated level is smaller than the depth value of the octree.
- a 3D cube area of nodes of a designated level according to embodiments is referred to as a block.
- One block may include one or more voxels.
- a block or voxel may correspond to a brick.
- geometry is represented as a surface.
- a surface according to embodiments may intersect each edge of a block at most once.
- intersection points there are at least 12 intersection points in one block. Each intersection point is called a vertex.
- a vertex existing along an edge is detected when there is at least one occupied voxel adjacent to the edge among all blocks sharing the edge.
- An occluded voxel according to embodiments means a voxel including a point.
- the position of a vertex detected along an edge is the average position along the edge of all voxels adjacent to the corresponding edge among all blocks sharing the corresponding edge.
- the point cloud encoder When a vertex is detected, the point cloud encoder according to the embodiments entropy-codes the starting point (x, y, z) of the edge, the direction vector ( ⁇ x, ⁇ y, ⁇ z) of the edge, and the vertex position value (relative position value within the edge).
- the point cloud encoder for example, the geometry reconstruction unit 40005
- the point cloud encoder performs triangle reconstruction, up-sampling, and voxelization processes. to create the restored geometry (reconstructed geometry).
- Vertices located at the edges of a block determine the surface through which the block passes.
- a surface according to embodiments is a non-planar polygon.
- the triangle reconstruction process reconstructs the surface represented by the triangle based on the starting point of the edge, the direction vector of the edge, and the position value of the vertex.
- the triangle reconstruction process is as follows. 1 Calculate the centroid value of each vertex, 2 Calculate the values obtained by subtracting the centroid value from each vertex value 3 Square the values, and obtain the sum of all the values.
- the minimum value of the added value is obtained, and the projection process is performed according to the axis with the minimum value. For example, if the x element is minimal, each vertex is projected along the x-axis based on the center of the block, and projected onto the (y, z) plane. If the value that results from projection on the (y, z) plane is (ai, bi), the ⁇ value is obtained through atan2(bi, ai), and the vertices are aligned based on the ⁇ value.
- the table below shows combinations of vertices to generate triangles according to the number of vertices. Vertices are sorted in order from 1 to n.
- the table below shows that two triangles can be formed for four vertices according to a combination of the vertices.
- the first triangle may be composed of the first, second, and third vertices among the aligned vertices
- the second triangle may be composed of the third, fourth, and first vertices among the aligned vertices. .
- the upsampling process is performed to voxelize by adding points in the middle along the edge of the triangle. Additional points are generated based on the upsampling factor and the width of the block. The added points are called refined vertices.
- a point cloud encoder may voxelize refined vertices. Also, the point cloud encoder may perform attribute encoding based on the voxelized position (or position value).
- FIG. 7 shows an example of a neighbor node pattern according to embodiments.
- the point cloud encoder may perform entropy coding based on context adaptive arithmetic coding.
- the point cloud content providing system or the point cloud encoder converts the occupancy code directly. Entropy coding is possible.
- the point cloud content providing system or the point cloud encoder performs entropy encoding (intra-encoding) based on the occupancy code of the current node and the occupancy of neighboring nodes, or entropy encoding (inter-encoding) based on the occupancy code of the previous frame. ) can be performed.
- a frame according to embodiments means a set of point cloud videos generated at the same time.
- Compression efficiency of intra-encoding/inter-encoding may vary according to the number of referenced neighboring nodes. If the bit size increases, it becomes complicated, but compression efficiency can be increased by making it skewed to one side. For example, if you have a 3-bit context, 2 of 3 should be coded in 8 ways. The part that is divided and coded affects the complexity of the implementation. Therefore, it is necessary to match the efficiency of compression with an appropriate level of complexity.
- a point cloud encoder determines occupancy of neighboring nodes of each node of an octree and obtains a neighboring node pattern value.
- the neighbor node pattern is used to infer the occupancy pattern of that node.
- the left side of FIG. 7 shows a cube corresponding to a node (a cube located in the middle) and six cubes (neighboring nodes) sharing at least one face with the cube.
- Nodes shown in the figure are nodes of the same depth (depth).
- the numbers shown in the figure represent weights (1, 2, 4, 8, 16, 32, etc.) associated with each of the six nodes. Each weight is sequentially assigned according to the locations of neighboring nodes.
- the right side of FIG. 7 shows neighboring node pattern values.
- the neighbor pattern value is the sum of values multiplied by the weights of the occupied neighbor nodes (neighbor nodes with points). Therefore, the neighbor node pattern values range from 0 to 63. If the neighbor node pattern value is 0, it indicates that there is no node (occupied node) having a point among the neighbor nodes of the corresponding node. When the neighbor node pattern value is 63, it indicates that all of the neighbor nodes are occupied nodes. As shown in the figure, since the neighboring nodes to which weights 1, 2, 4, and 8 are assigned are ocupied nodes, the neighboring node pattern value is 15, which is the value obtained by adding 1, 2, 4, and 8.
- the point cloud encoder may perform coding according to the neighboring node pattern value (for example, if the neighboring node pattern value is 63, 64 types of coding are performed). According to embodiments, the point cloud encoder may reduce complexity of coding by changing a neighbor node pattern value (for example, based on a table changing 64 to 10 or 6).
- the encoded geometry is reconstructed (decompressed) before attribute encoding is performed.
- the geometry reconstruction operation may include changing the arrangement of direct coded points (eg, placing the direct coded points in front of point cloud data).
- the geometry reconstruction process includes triangle reconstruction, upsampling, and voxelization processes. Since attributes depend on the geometry, attribute encoding is performed based on the reconstructed geometry.
- the point cloud encoder may reorganize points according to LODs.
- the diagram shows point cloud content corresponding to the LOD.
- the left side of the figure shows the original point cloud content.
- the second figure from the left of the figure shows the distribution of points with the lowest LOD, and the rightmost figure of the figure shows the distribution of points with the highest LOD. That is, points of the lowest LOD are sparsely distributed, and points of the highest LOD are densely distributed. That is, as the LOD increases according to the direction of the arrow indicated at the bottom of the drawing, the interval (or distance) between points becomes shorter.
- the point cloud content providing system or the point cloud encoder (eg, the point cloud video encoder 10002, the point cloud encoder in FIG. 4, or the LOD generator 40009) generates an LOD. can do.
- the LOD is created by reorganizing the points into a set of refinement levels according to a set LOD distance value (or set of Euclidean Distances).
- the LOD generation process is performed in the point cloud decoder as well as the point cloud encoder.
- FIG. 9 shows examples of points (P0 to P9) of the point cloud content distributed in the 3D space.
- the original order of FIG. 9 indicates the order of points P0 to P9 before LOD generation.
- the LOD based order of FIG. 9 represents the order of points according to LOD generation. Points are reordered by LOD. Also, high LOD includes points belonging to low LOD.
- LOD0 includes P0, P5, P4, and P2.
- LOD1 contains the points of LOD0 and P1, P6 and P3.
- LOD2 includes points of LOD0, points of LOD1 and P9, P8 and P7.
- the point cloud encoder may perform prediction transform coding, lifting transform coding, and RAHT transform coding selectively or in combination.
- a point cloud encoder may generate predictors for points and perform predictive transformation coding to set predictive attributes (or predictive attribute values) of each point. That is, N predictors can be generated for N points.
- the predicted attribute (or attribute value) is a weight calculated based on the distance to each neighboring point to the attributes (or attribute values, eg, color, reflectance, etc.) of neighboring points set in the predictor of each point. (or weight value) is set as the average value of multiplied values.
- the point cloud encoder for example, the coefficient quantization unit 40011 according to the embodiments subtracts the predicted attribute (attribute value) from the attribute (attribute value) of each point, and generates residual values (residuals, residual attributes, residual attribute values, attributes) can be called a prediction residual, etc.) can be quatized and inverse quantized, and the quantization process is shown in the following table.
- the point cloud encoder (for example, the Arismetic encoder 40012) according to the embodiments may entropy code the quantized and inverse quantized residual values as described above when there are points adjacent to the predictor of each point.
- the point cloud encoder (for example, the Arismetic encoder 40012) according to the examples may entropy code attributes of the corresponding point without performing the above-described process if there are no neighboring points to the predictor of each point.
- the point cloud encoder (for example, the lifting transform unit 40010) according to the embodiments generates a predictor of each point, sets the LOD calculated in the predictor, registers neighboring points, and weights according to distances to neighboring points.
- Lifting transform coding according to the embodiments is similar to the above-described predictive transform coding, but is different in that weights are cumulatively applied to attribute values. The process of cumulatively applying weights to values is as follows.
- the weight calculated for all predictors is additionally multiplied by the weight stored in the QW corresponding to the predictor index, and the calculated weight is cumulatively summed as the index of the neighboring node in the update weight array.
- the value obtained by multiplying the calculated weight by the attribute value of the index of the neighboring node is cumulatively summed.
- a predicted attribute value is calculated by additionally multiplying an attribute value updated through the lift update process by a weight updated through the lift prediction process (stored in QW).
- a point cloud encoder eg, the coefficient quantization unit 40011
- the point cloud encoder eg, the Arismetic encoder 40012
- the point cloud encoder (for example, the RAHT transform unit 40008) according to the embodiments may perform RAHT transform coding to predict attributes of nodes at a higher level using attributes associated with nodes at a lower level of the octree. .
- RAHT transform coding is an example of attribute intra coding through octree backward scan.
- the point cloud encoder according to embodiments scans from voxels to the entire area and repeats the merging process up to the root node while merging the voxels into larger blocks in each step.
- a merging process according to embodiments is performed only for Occupied nodes.
- a merging process is not performed on an empty node, but a merging process is performed on an immediate parent node of an empty node.
- the following equation represents a RAHT transformation matrix.
- g lx, y, z represent average attribute values of voxels at level l.
- g lx, y, z can be computed from g l+1 2x, y, z and g l+1 2x+1, y, z .
- g l-1 x, y, z are low-pass values and are used in the merging process at the next higher level.
- h l ⁇ 1 x, y, and z are high-pass coefficients, and the high-pass coefficients at each step are quantized and entropy-coded (for example, encoding of the Arithmetic Encoder 400012).
- the root node is created as follows through the last g 1 0, 0, 0 and g 1 0, 0, 1 ,
- FIG. 10 shows an example of a point cloud decoder according to embodiments.
- the point cloud decoder shown in FIG. 10 is an example of the point cloud video decoder 10006 described in FIG. 1 , and may perform the same or similar operation as the operation of the point cloud video decoder 10006 described in FIG. 1 .
- the point cloud decoder may receive a geometry bitstream and an attribute bitstream included in one or more bitstreams.
- the point cloud decoder includes a geometry decoder and an attribute decoder.
- the geometry decoder performs geometry decoding on a geometry bitstream and outputs decoded geometry.
- the attribute decoder performs attribute decoding based on the decoded geometry and attribute bitstream and outputs decoded attributes.
- the decoded geometry and decoded attributes are used to reconstruct the point cloud content (decoded point cloud).
- FIG. 11 shows an example of a point cloud decoder according to embodiments.
- the point cloud decoder shown in FIG. 11 is an example of the point cloud decoder described in FIG. 10 and can perform a decoding operation, which is the reverse process of the encoding operation of the point cloud encoder described in FIGS. 1 to 9 .
- the point cloud decoder may perform geometry decoding and attribute decoding. Geometry decoding is performed before attribute decoding.
- the point cloud decoder includes an arithmetic decoder (11000), an octree synthesizer (synthesize octree) 11001, a surface synthesize surface approximation unit (11002), and a geometry reconstructor (reconstruct geometry). . ), an inverse lifting unit (11009), and/or an inverse transform colors (11010).
- the Arismetic decoder 11000, the octree synthesizer 11001, the surface deoxymation synthesizer 11002, the geometry reconstructor 11003, and the coordinate system inverse transform unit 11004 may perform geometry decoding.
- Geometry decoding according to embodiments may include direct coding and trisoup geometry decoding. Direct coding and tri-sup geometry decoding are selectively applied. Also, geometry decoding is not limited to the above example, and is performed in a reverse process to the geometry encoding described in FIGS. 1 to 9 .
- the Arismetic decoder 11000 decodes the received geometry bitstream based on Arithmetic coding.
- the operation of the Arithmetic Decoder 11000 corresponds to the reverse process of the Arithmetic Encoder 40004.
- the octree synthesizer 11001 may generate an octree by obtaining an occupancy code from a decoded geometry bitstream (or information on geometry obtained as a result of decoding). A detailed description of the occupancy code is as described with reference to FIGS. 1 to 9 .
- the surface deoxymation synthesis unit 11002 may synthesize a surface based on the decoded geometry and/or the generated octree.
- the geometry reconstructor 11003 may regenerate geometry based on surfaces and/or decoded geometry. As described in FIGS. 1 to 9 , direct coding and tri-sup geometry encoding are selectively applied. Accordingly, the geometry reconstruction unit 11003 directly imports and adds position information of points to which direct coding is applied. In addition, when triangle geometry encoding is applied, the geometry reconstructor 11003 may perform reconstruction operations of the geometry reconstructor 40005, for example, triangle reconstruction, up-sampling, and voxelization operations to restore the geometry. there is. Details are the same as those described in FIG. 6 and thus are omitted.
- the reconstructed geometry may include a point cloud picture or frame that does not include attributes.
- the coordinate system inverse transformation unit 11004 may obtain positions of points by transforming the coordinate system based on the restored geometry.
- the Arithmetic Decoder 11005, Inverse Quantization Unit 11006, RAHT Transformation Unit 11007, LOD Generator 11008, Inverse Lifting Unit 11009, and/or Color Inverse Transformation Unit 11010 are the attributes described with reference to FIG. decoding can be performed.
- Attribute decoding according to embodiments includes Region Adaptive Hierarchial Transform (RAHT) decoding, Interpolaration-based hierarchical nearest-neighbour prediction-Prediction Transform decoding, and interpolation-based hierarchical nearest-neighbour prediction with an update/lifting transform. step (Lifting Transform)) decoding.
- RAHT Region Adaptive Hierarchial Transform
- Interpolaration-based hierarchical nearest-neighbour prediction-Prediction Transform decoding and interpolation-based hierarchical nearest-neighbour prediction with an update/lifting transform.
- step (Lifting Transform)) decoding The above three decodings may be selectively used, or a combination of one or more decodings may
- the Arismetic decoder 11005 decodes the attribute bitstream by Arithmetic coding.
- the inverse quantization unit 11006 inverse quantizes the decoded attribute bitstream or information about attributes obtained as a result of decoding, and outputs inverse quantized attributes (or attribute values). Inverse quantization may be selectively applied based on attribute encoding of the point cloud encoder.
- the RAHT conversion unit 11007, the LOD generation unit 11008, and/or the inverse lifting unit 11009 may process the reconstructed geometry and inverse quantized attributes. As described above, the RAHT converter 11007, the LOD generator 11008, and/or the inverse lifter 11009 may selectively perform a decoding operation corresponding to the encoding of the point cloud encoder.
- the color inverse transform unit 11010 performs inverse transform coding for inverse transform of color values (or textures) included in decoded attributes.
- the operation of the inverse color transform unit 11010 may be selectively performed based on the operation of the color transform unit 40006 of the point cloud encoder.
- elements of the point cloud decoder of FIG. 11 are not shown in the figure, hardware including one or more processors or integrated circuits configured to communicate with one or more memories included in the point cloud providing device. , may be implemented in software, firmware, or a combination thereof. One or more processors may perform at least one or more of the operations and/or functions of the elements of the point cloud decoder of FIG. 11 described above. Also, one or more processors may operate or execute a set of software programs and/or instructions to perform operations and/or functions of elements of the point cloud decoder of FIG. 11 .
- the transmission device shown in FIG. 12 is an example of the transmission device 10000 of FIG. 1 (or the point cloud encoder of FIG. 4 ).
- the transmission device shown in FIG. 12 may perform at least one or more of operations and methods identical or similar to the operations and encoding methods of the point cloud encoder described in FIGS. 1 to 9 .
- a transmission device includes a data input unit 12000, a quantization processing unit 12001, a voxelization processing unit 12002, an octree occupancy code generation unit 12003, a surface model processing unit 12004, an intra/ Inter-coding processing unit 12005, Arithmetic coder 12006, metadata processing unit 12007, color conversion processing unit 12008, attribute conversion processing unit (or attribute conversion processing unit) 12009, prediction/lifting/RAHT conversion It may include a processing unit 12010, an Arithmetic coder 12011 and/or a transmission processing unit 12012.
- the data input unit 12000 receives or acquires point cloud data.
- the data input unit 12000 may perform the same or similar operation and/or acquisition method to the operation and/or acquisition method of the point cloud video acquisition unit 10001 (or the acquisition process 20000 described in FIG. 2 ).
- Geometry encoding according to embodiments is the same as or similar to the geometry encoding described with reference to FIGS. 1 to 9, and thus a detailed description thereof will be omitted.
- the quantization processor 12001 quantizes geometry (eg, position values or position values of points).
- the operation and/or quantization of the quantization processing unit 12001 is the same as or similar to the operation and/or quantization of the quantization unit 40001 described with reference to FIG. 4 .
- a detailed description is the same as that described in FIGS. 1 to 9 .
- the voxelization processor 12002 voxelizes position values of quantized points.
- the voxelization processing unit 120002 may perform the same or similar operations and/or processes to those of the quantization unit 40001 described with reference to FIG. 4 and/or the voxelization process. A detailed description is the same as that described in FIGS. 1 to 9 .
- the octree occupancy code generation unit 12003 performs octree coding on positions of voxelized points based on an octree structure.
- the octree occupancy code generator 12003 may generate an occupancy code.
- the octree occupancy code generator 12003 may perform operations and/or methods identical or similar to those of the point cloud encoder (or the octree analyzer 40002) described with reference to FIGS. 4 and 6 . A detailed description is the same as that described in FIGS. 1 to 9 .
- the surface model processing unit 12004 may perform tri-sup geometry encoding to reconstruct positions of points within a specific area (or node) based on a surface model on a voxel basis.
- the four-surface model processing unit 12004 may perform the same or similar operation and/or method to the operation and/or method of the point cloud encoder (eg, the surface approximation analysis unit 40003) described with reference to FIG. 4 .
- a detailed description is the same as that described in FIGS. 1 to 9 .
- the intra/inter coding processing unit 12005 may intra/inter code the point cloud data.
- the intra/inter coding processing unit 12005 may perform coding identical to or similar to the intra/inter coding described with reference to FIG. 7 . A detailed description is the same as that described in FIG. 7 .
- the intra/inter coding processor 12005 may be included in the Arithmetic Coder 12006.
- Arithmetic coder 12006 entropy encodes an octree of point cloud data and/or an approximated octree.
- the encoding method includes an Arithmetic encoding method.
- Arithmetic coder 12006 performs the same or similar operations and/or methods to operations and/or methods of Arithmetic encoder 40004.
- the metadata processing unit 12007 processes metadata about point cloud data, for example, set values, and provides them to a necessary process such as geometry encoding and/or attribute encoding. Also, the metadata processing unit 12007 according to embodiments may generate and/or process signaling information related to geometry encoding and/or attribute encoding. Signaling information according to embodiments may be encoded separately from geometry encoding and/or attribute encoding. Also, signaling information according to embodiments may be interleaved.
- a color conversion processing unit 12008, an attribute conversion processing unit 12009, a prediction/lifting/RAHT conversion processing unit 12010, and an Arithmetic coder 12011 perform attribute encoding.
- Attribute encoding according to embodiments is the same as or similar to the attribute encoding described with reference to FIGS. 1 to 9, so a detailed description thereof will be omitted.
- the color conversion processing unit 12008 performs color conversion coding to convert color values included in attributes.
- the color conversion processing unit 12008 may perform color conversion coding based on the reconstructed geometry. Description of the reconstructed geometry is the same as that described in FIGS. 1 to 9 . In addition, the same or similar operations and/or methods to those of the color conversion unit 40006 described in FIG. 4 are performed. A detailed description is omitted.
- the attribute transformation processing unit 12009 performs attribute transformation to transform attributes based on positions for which geometry encoding has not been performed and/or reconstructed geometry.
- the attribute conversion processing unit 12009 performs the same or similar operation and/or method to the operation and/or method of the attribute conversion unit 40007 described in FIG. 4 .
- a detailed description is omitted.
- the prediction/lifting/RAHT transform processing unit 12010 may code the transformed attributes with any one or combination of RAHT coding, prediction transform coding, and lifting transform coding.
- the prediction/lifting/RAHT conversion processing unit 12010 performs at least one of the same or similar operations to those of the RAHT conversion unit 40008, the LOD generation unit 40009, and the lifting conversion unit 40010 described in FIG. 4 do.
- descriptions of predictive transform coding, lifting transform coding, and RAHT transform coding are the same as those described in FIGS. 1 to 9, so detailed descriptions thereof are omitted.
- the Arithmetic Coder 12011 may encode coded attributes based on Arithmetic Coding.
- the Arithmetic Coder 12011 performs the same or similar operations and/or methods to those of the Arithmetic Encoder 400012.
- the transmission processing unit 12012 transmits each bitstream including encoded geometry and/or encoded attributes and metadata information, or transmits encoded geometry and/or encoded attributes and metadata information into one It can be configured as a bitstream and transmitted.
- the bitstream may include one or more sub-bitstreams.
- the bitstream according to the embodiments includes Sequence Parameter Set (SPS) for signaling at the sequence level, Geometry Parameter Set (GPS) for signaling of geometry information coding, Attribute Parameter Set (APS) for signaling of attribute information coding, tile It may include signaling information and slice data including TPS (Tile Parameter Set) for level signaling.
- Slice data may include information on one or more slices.
- One slice according to embodiments may include one geometry bitstream Geom00 and one or more attribute bitstreams Attr00 and Attr10.
- a slice refers to a series of syntax elements representing all or part of a coded point cloud frame.
- a TPS may include information about each tile (for example, coordinate value information and height/size information of a bounding box) for one or more tiles.
- a geometry bitstream may include a header and a payload.
- the header of the geometry bitstream may include identification information (geom_parameter_set_id) of a parameter set included in GPS, a tile identifier (geom_tile_id), a slice identifier (geom_slice_id), and information about data included in a payload.
- the metadata processing unit 12007 may generate and/or process signaling information and transmit it to the transmission processing unit 12012.
- elements performing geometry encoding and elements performing attribute encoding may share data/information with each other as indicated by dotted lines.
- the transmission processing unit 12012 may perform the same or similar operation and/or transmission method to the operation and/or transmission method of the transmitter 10003. A detailed description is omitted since it is the same as that described in FIGS. 1 and 2 .
- FIG. 13 is an example of a receiving device according to embodiments.
- the receiving device shown in FIG. 13 is an example of the receiving device 10004 of FIG. 1 (or the point cloud decoder of FIGS. 10 and 11).
- the receiving device illustrated in FIG. 13 may perform at least one or more of operations and methods identical or similar to the operations and decoding methods of the point cloud decoder described in FIGS. 1 to 11 .
- a receiving device includes a receiving unit 13000, a receiving processing unit 13001, an arithmetic decoder 13002, an octree reconstruction processing unit 13003 based on an occupancy code, and a surface model processing unit (triangle reconstruction). , up-sampling, voxelization) 13004, inverse quantization processing unit 13005, metadata parser 13006, arithmetic decoder 13007, inverse quantization processing unit 13008, prediction It may include a /lifting/RAHT inverse transformation processing unit 13009, a color inverse transformation processing unit 13010, and/or a renderer 13011.
- Each component of decoding according to the embodiments may perform a reverse process of the component of encoding according to the embodiments.
- the receiving unit 13000 receives point cloud data.
- the receiver 13000 may perform the same or similar operation and/or reception method to the operation and/or reception method of the receiver 10005 of FIG. 1 . A detailed description is omitted.
- the reception processing unit 13001 may obtain a geometry bitstream and/or an attribute bitstream from received data.
- the receiving processing unit 13001 may be included in the receiving unit 13000.
- the Arismetic decoder 13002, the octree reconstruction processing unit 13003 based on the occupancy code, the surface model processing unit 13004, and the inverse quantization processing unit 13005 may perform geometry decoding.
- Geometry decoding according to the embodiments is the same as or similar to the geometry decoding described in FIGS. 1 to 10, and thus a detailed description thereof will be omitted.
- the Arismetic decoder 13002 may decode a geometry bitstream based on Arithmetic coding.
- the Arismetic decoder 13002 performs the same or similar operation and/or coding to that of the Arithmetic decoder 11000.
- the octree reconstruction processing unit 13003 based on occupancy code may obtain an occupancy code from a decoded geometry bitstream (or information about a geometry secured as a result of decoding) to reconstruct an octree.
- the octree reconstruction processing unit 13003 based on the occupancy code performs the same or similar operations and/or methods to those of the octree synthesis unit 11001 and/or the octree generation method.
- the surface model processing unit 13004 according to embodiments performs tri-soup geometry decoding based on the surface model method and related geometry reconstruction (eg, triangle reconstruction, up-sampling, and voxelization) when tri-sup geometry encoding is applied. can be performed.
- the surface model processing unit 13004 performs operations identical to or similar to those of the surface deoxymation synthesis unit 11002 and/or the geometry reconstruction unit 11003.
- the inverse quantization processor 13005 may inverse quantize the decoded geometry.
- the metadata parser 13006 may parse metadata included in the received point cloud data, for example, setting values. Metadata parser 13006 can pass metadata to geometry decoding and/or attribute decoding. A detailed description of the metadata is omitted since it is the same as that described in FIG. 12 .
- the Arismetic decoder 13007, the inverse quantization processing unit 13008, the prediction/lifting/RAHT inverse transformation processing unit 13009, and the color inverse transformation processing unit 13010 perform attribute decoding. Attribute decoding is the same as or similar to the attribute decoding described in FIGS. 1 to 10, so a detailed description thereof will be omitted.
- the Arismetic decoder 13007 may decode the attribute bitstream through Arismetic coding.
- the Arismetic decoder 13007 may perform decoding of the attribute bitstream based on the reconstructed geometry.
- the Arismetic decoder 13007 performs the same or similar operation and/or coding to that of the Arithmetic decoder 11005.
- the inverse quantization processing unit 13008 may inverse quantize the decoded attribute bitstream.
- the inverse quantization processing unit 13008 performs the same or similar operation and/or method to the operation and/or inverse quantization method of the inverse quantization unit 11006.
- the prediction/lifting/RAHT inverse transform processing unit 13009 may process reconstructed geometry and inverse quantized attributes.
- the prediction/lifting/RAHT inverse transform processing unit 13009 performs operations identical or similar to those of the RAHT transform unit 11007, the LOD generator 11008 and/or the inverse lifting unit 11009 and/or decoding operations and/or At least one or more of decoding is performed.
- the inverse color transformation processing unit 13010 according to embodiments performs inverse transformation coding for inversely transforming color values (or textures) included in decoded attributes.
- the inverse color transform processing unit 13010 performs the same or similar operation and/or inverse transform coding to the operation and/or inverse transform coding of the inverse color transform unit 11010.
- the renderer 13011 may render point cloud data.
- FIG. 14 shows an example of a structure capable of interworking with a method/apparatus for transmitting and receiving point cloud data according to embodiments.
- the structure of FIG. 14 includes at least one of a server 1460, a robot 1410, an autonomous vehicle 1420, an XR device 1430, a smartphone 1440, a home appliance 1450, and/or an HMD 1470. It shows a configuration connected to the cloud network 1410.
- a robot 1410, an autonomous vehicle 1420, an XR device 1430, a smartphone 1440 or a home appliance 1450 are referred to as devices.
- the XR device 1430 may correspond to or interwork with a point cloud data (PCC) device according to embodiments.
- PCC point cloud data
- the cloud network 1400 may constitute a part of a cloud computing infrastructure or may refer to a network existing in a cloud computing infrastructure.
- the cloud network 1400 may be configured using a 3G network, a 4G or Long Term Evolution (LTE) network, or a 5G network.
- LTE Long Term Evolution
- the server 1460 connects at least one of the robot 1410, the autonomous vehicle 1420, the XR device 1430, the smartphone 1440, the home appliance 1450, and/or the HMD 1470 to the cloud network 1400. It is connected through and may help at least part of the processing of the connected devices 1410 to 1470.
- a Head-Mount Display (HMD) 1470 represents one of types in which an XR device and/or a PCC device according to embodiments may be implemented.
- An HMD type device includes a communication unit, a control unit, a memory unit, an I/O unit, a sensor unit, and a power supply unit.
- devices 1410 to 1450 to which the above-described technology is applied will be described.
- the devices 1410 to 1450 shown in FIG. 14 may interwork/combine with the device for transmitting/receiving point cloud data according to the above-described embodiments.
- the XR/PCC device 1430 applies PCC and/or XR (AR+VR) technology to a Head-Mount Display (HMD), a Head-Up Display (HUD) installed in a vehicle, a television, a mobile phone, a smart phone, It may be implemented as a computer, a wearable device, a home appliance, a digital signage, a vehicle, a fixed robot or a mobile robot.
- HMD Head-Mount Display
- HUD Head-Up Display
- the XR/PCC device 1430 analyzes 3D point cloud data or image data obtained through various sensors or from an external device to generate positional data and attribute data for 3D points, thereby generating positional data and attribute data for surrounding space or real objects. Information can be obtained, and XR objects to be displayed can be rendered and output. For example, the XR/PCC device 1430 may output an XR object including additional information about the recognized object in correspondence with the recognized object.
- the XR/PCC device 1430 may be implemented as a mobile phone 1440 or the like to which PCC technology is applied.
- the mobile phone 1440 may decode and display point cloud content based on PCC technology.
- the self-driving vehicle 1420 may be implemented as a mobile robot, vehicle, unmanned aerial vehicle, etc. by applying PCC technology and XR technology.
- the self-driving vehicle 1420 to which XR/PCC technology is applied may refer to an autonomous vehicle equipped with a means for providing XR images or an autonomous vehicle subject to control/interaction within the XR images.
- the self-driving vehicle 1420 which is a target of control/interaction within the XR image, is distinguished from the XR device 1430 and may be interlocked with each other.
- the self-driving vehicle 1420 equipped with a means for providing an XR/PCC image may obtain sensor information from sensors including cameras and output an XR/PCC image generated based on the obtained sensor information.
- the self-driving vehicle 1420 may provide an XR/PCC object corresponding to a real object or an object in a screen to a passenger by outputting an XR/PCC image with a HUD.
- the XR/PCC object when the XR/PCC object is output to the HUD, at least a part of the XR/PCC object may be output to overlap the real object toward which the passenger's gaze is directed.
- an XR/PCC object when an XR/PCC object is output to a display provided inside an autonomous vehicle, at least a part of the XR/PCC object may be output to overlap the object in the screen.
- the autonomous vehicle 1220 may output XR/PCC objects corresponding to objects such as lanes, other vehicles, traffic lights, traffic signs, two-wheeled vehicles, pedestrians, and buildings.
- VR Virtual Reality
- AR Augmented Reality
- MR Mixed Reality
- PCC Point Cloud Compression
- VR technology is a display technology that provides objects or backgrounds of the real world only as CG images.
- AR technology means a technology that shows a virtually created CG image on top of a real object image.
- MR technology is similar to the aforementioned AR technology in that it mixes and combines virtual objects in the real world.
- real objects and virtual objects made of CG images are clear, and virtual objects are used in a form that complements real objects, whereas in MR technology, virtual objects are considered equivalent to real objects. distinct from technology. More specifically, for example, a hologram service to which the above-described MR technology is applied.
- VR, AR, and MR technologies are sometimes referred to as XR (extended reality) technologies rather than clearly distinguishing them. Accordingly, embodiments of the present invention are applicable to all VR, AR, MR, and XR technologies. As for this technique, encoding/decoding based on PCC, V-PCC, and G-PCC techniques may be applied.
- the PCC method/apparatus according to the embodiments may be applied to vehicles providing autonomous driving services.
- a vehicle providing autonomous driving service is connected to a PCC device to enable wired/wireless communication.
- Point cloud data (PCC) transmission and reception devices when connected to enable wired/wireless communication with a vehicle, receive/process content data related to AR/VR/PCC services that can be provided together with autonomous driving services to provide a vehicle can be sent to
- the point cloud transmission/reception device when the point cloud data transmission/reception device is mounted on a vehicle, the point cloud transmission/reception device may receive/process AR/VR/PCC service-related content data according to a user input signal input through a user interface device and provide the received/processed content data to the user.
- a vehicle or user interface device may receive a user input signal.
- a user input signal according to embodiments may include a signal indicating an autonomous driving service.
- the point cloud data transmission method/device includes the transmission device 10000 of FIG. 1, the point cloud video encoder 10002, the transmitter 10003, and the acquisition-encoding-transmission (20000-20001-20002) of FIG. , encoder of FIG. 4, transmitter of FIG. 12, device of FIG. 14, encoding of FIGS. 15 to 23, encoding of FIGS. 27 to 29, encoding of FIGS. 33 to 36, bitstream and parameter generation, FIG. 37, encoding of FIGS. 40-42, FIG. 43 It is interpreted as a term indicating the transmission method, etc.
- a method/device for receiving point cloud data includes a receiving device 10004 of FIG. 1, a receiver 10005, a point cloud video decoder 10006, and transmission-decoding-rendering (20002-20003-20004) of FIG. , decoder of FIGS. 10-11, receiver of FIG. 13, device of FIG. 14, decoding of FIGS. 15 to 23, decoding of FIGS. 27 to 29, parsing of the bitstream of FIGS. 33 to 36, decoding of FIG. 38, decoding of FIGS. 40-42 , It is interpreted as a term referring to the FIG. 44 receiving method and the like.
- the method/device for transmitting/receiving point cloud data may be referred to as a method/device according to embodiments.
- geometry data, geometry information, location information, etc. constituting point cloud data are interpreted as the same meaning.
- Attribute data, attribute information, and attribute information constituting the point cloud data are interpreted as the same meaning.
- a method/apparatus may provide a method for controlling point cloud density in a slice.
- Embodiments provide a method for efficiently supporting selective decoding of a portion of data due to receiver performance or transmission speed when transmitting and receiving point cloud data.
- geometry and attribute data transmitted in data units are divided into semantic units such as geometry octree and level of detail (LoD), which is necessary in bitstream units. You can select the information to be or remove unnecessary information.
- Embodiments address techniques for constructing data structures composed of point clouds. Specifically, a packing and signaling method for effectively delivering layer-based PCC data is described, and based on this, a method for applying to a scalable PCC-based service is proposed. In particular, when the direct compression mode is used for positional compression, a method for constructing and transmitting/receiving a slice segment to be more suitable for a scalable PCC service is proposed. In particular, we propose a compression structure for efficient storage and transmission of large-capacity point cloud data with a wide distribution and high point density.
- point cloud data is composed of a location (geometry: e.g., XYZ coordinates) and attributes (e.g., color, reflectance, intensity, grayscale, opacity, etc.) of each data.
- attributes e.g., color, reflectance, intensity, grayscale, opacity, etc.
- octree-based compression is performed to efficiently compress distribution characteristics that are non-uniformly distributed in a 3-dimensional space, and attribute information is compressed based on this.
- FIGS. 4 and 11 point cloud data is composed of a location (geometry: e.g., XYZ coordinates) and attributes (e.g., color, reflectance, intensity, grayscale, opacity, etc.) of each data.
- PCC Point Cloud Compression
- octree-based compression is performed to efficiently compress distribution characteristics that are non-uniformly distributed in a 3-dimensional space, and attribute information is compressed based on this.
- 15 illustrates a process of encoding, transmitting, and decoding point cloud data according to embodiments.
- Each component of FIG. 15 may correspond to hardware, software, processor, and/or a combination thereof.
- the point cloud encoder 15000 is a transmission device according to embodiments that performs a transmission method according to embodiments, and can scalably encode and transmit point cloud data.
- the point cloud decoder 15010 is a receiving device according to embodiments that performs a receiving method according to embodiments, and can scalably decode point cloud data.
- Source data received by the encoder 15000 may include geometry data and/or attribute data.
- the encoder 15000 scalably encodes the point cloud data and does not directly generate a partial PCC bitstream, receives full geometry data and full attribute data, stores the data in storage connected to the encoder, and then For partial encoding, transcoding may be performed to generate and transmit a partial PCC bitstream.
- the decoder 15010 may receive and decode the partial PCC bitstream to restore partial geometry and/or partial attributes.
- the encoder 15000 may receive the full geometry and full attributes, store the data in storage connected to the encoder, transcode the point cloud data with a low QP (quantization parameter), and generate and transmit the entire PCC bitstream.
- the decoder 15010 may receive and decode the entire PCC bitstream to recover full geometry and/or full attributes.
- the decoder 15010 may select partial geometry and/or partial attributes from the entire PCC bitstream through data selection.
- the method/device divides location information and characteristic information such as color/brightness/reflectivity of a data point, which is point cloud data, into geometry and attribute information, compresses and transmits them, respectively.
- PCC data may be configured according to an octree structure having layers or a level of detail (LoD) according to a degree of detail. Based on this, scalable point cloud data coding and representation are possible. At this time, it is possible to decode or represent only a part of the point cloud data according to the performance or transmission speed of the receiver.
- the method/device according to the embodiments may remove unnecessary data in advance in this process. That is, when only a part of the scalable PCC bitstream needs to be transmitted (when only some layers are decoded during scalable decoding), only the required part is selected and transmitted. Since it cannot be done, 1) re-encode the necessary part after decoding (15020) or 2) transmit the whole and then selectively apply it in the receiver (15030). However, in the case of 1), a delay may occur due to the time for decoding and re-encoding (15020), and in the case of 2), bandwidth efficiency decreases due to transmission of unnecessary data, and a fixed bandwidth ), there is a point in transmitting with lower data quality (15030).
- a method/device may define a slice subdivision structure of point cloud data and signal a scalable layer and a slice structure for scalable transmission.
- Embodiments may classify and process bitstreams in specific units for efficient bitstream delivery and decoding.
- an entropy-based compression method and direct coding may be used together in the case of octree-based location compression.
- a slice configuration is required for
- Units according to embodiments may be referred to as LOD, layer, slice, and the like.
- LOD is the same term as LOD of attribute data coding, but may mean a data unit for a layer structure of a bitstream in another meaning. It may be a concept of corresponding to one depth based on a hierarchical structure of point cloud data, eg, depth (level) of an octree or several trees, or combining two or more depths.
- a layer is for generating a unit of a sub-bitstream, corresponds to one depth or is a concept of combining two or more depths, and may correspond to one LOD or two or more LODs.
- a slice is a unit for constructing a unit of a sub-bitstream, and may correspond to one depth, a part of one depth, or two or more depths. Also, whether it corresponds to one LOD or a part of one LOD, it may correspond to two or more LODs.
- LODs, layers, and slices may correspond to each other or have an inclusive relationship. Also, units according to embodiments include LODs, layers, slices, layer groups, subgroups, and the like, and may be referred to interchangeably.
- the embodiments include a region-of-interest-based slice selective coding and transmission method, a slice generation method considering the density of cloud data, a fixed subgroup size, a fixed subgroup size, and a subgroup size. It may include a group division (Fixed subgroupSize + subgroup division), a rotation-based subgroup division (subgroup division) signaling method, and the like.
- FIG. 16 illustrates a layer-based point cloud data configuration according to embodiments and a geometry and attribute bitstream structure according to embodiments.
- a transmission method/apparatus may configure layer-based point cloud data as shown in FIG. 16 to encode and decode the point cloud data.
- Embodiments aim at efficient transmission and decoding by selectively transmitting and decoding data in a bitstream unit for point cloud data composed of layers.
- Layering of point cloud data is layered from various viewpoints such as SNR, spatial resolution, color, temporal frequency, and bit depth, depending on the application field.
- layers may be formed in a direction in which the density of data increases.
- a method/apparatus may configure, encode, and decode a geometry bitstream and an attribute bitstream based on layering as shown in FIG. 16 .
- a bitstream obtained through point cloud compression of a transmission device/encoder is divided into a geometry data bitstream and an attribute data bitstream according to the type of data (attribute data bitstream) and can be transmitted.
- Each bitstream according to embodiments may be configured as a slice and transmitted. Regardless of layer information or LoD information, a geometry data bitstream and an attribute data bitstream may be configured as one slice and transmitted. In this case, if the layer ( 1) Process of decoding bitstream 2) Process of selecting only the part to be used and removing unnecessary part 3) Re-encoding based on only necessary information encoding) process.
- 17 shows a bitstream configuration according to embodiments.
- a transmitting method/device may generate a bitstream as shown in FIG. 17, and a receiving method/device according to embodiments may decode point cloud data included in the bitstream as shown in FIG. 17.
- Embodiments may apply a method of dividing and transmitting bitstreams in units of layers (or LoD) to avoid unnecessary intermediate processes.
- a low LoD is included in a high LoD.
- Information included in the current LoD but not included in the previous LoD, that is, information newly included for each LoD may be referred to as R (rest).
- R information newly included for each LoD.
- initial LoD information and information R newly included in each LoD can be divided into independent units and transmitted.
- a transmission method/apparatus may encode geometry data and generate a geometry bitstream.
- the geometry bitstream may be configured for each LOD or layer, and the geometry bitstream may include a header (geometry header) for each LOD or layer configuration unit.
- the header may include reference information for the next LOD or next layer.
- the current LOD (layer) may further include R information (geometry data) not included in the previous LOD (layer).
- a receiving method/device may encode attribute data and generate an attribute bitstream.
- the attribute bitstream may be configured for each LOD or layer, and the attribute bitstream may include a header (attribute header) for each LOD or layer.
- the header may include reference information for the next LOD or next layer.
- the current LOD (layer) may further include R information (attribute data) not included in the previous LOD (layer).
- the receiving method/device according to the embodiments may receive a bitstream composed of LODs or layers and efficiently decode only data to be used without a complicated intermediate process.
- a method/apparatus according to embodiments may align the bitstream of FIG. 17 as shown in FIG. 18.
- a transmission method/apparatus may serially transmit geometry and attributes as shown in FIG. 18 when transmitting a bitstream.
- the entire geometry information may be transmitted first according to the type of data, and then the attribute information (attribute data) may be transmitted.
- attribute information attribute data
- layers (LODs) including geometry data may be located first in the bitstream, and layers (LODs) including attribute data may be located after the geometry layer. Since attribute data depends on geometry data, the geometry layer can be placed first. In addition, the position can be changed in various ways according to embodiments. References between geometry headers are possible, and references between attribute headers and geometry headers are also possible.
- bitstreams constituting the same layer including geometry data and attribute data may be collected and transmitted.
- the decoding execution time can be shortened.
- information that needs to be processed first small LoD, geometry should precede attributes
- the first layer 1800 includes geometry data and attribute data corresponding to the smallest LOD 0 (layer 0) along with each header, and the second layer 1810 includes LOD 0 (layer 0). Geometry data and attribute data of points for new and more detailed layer 1 (LOD 1) not in LOD 0 (layer 0) are included as R1 information. Similarly, a third layer 1820 may follow.
- LOD 1 new and more detailed layer 1
- the transmission/reception method/device can efficiently select a layer (or LoD) desired in an application field at a bitstream level when transmitting and receiving a bitstream.
- a layer or LoD
- bitstream alignment methods when geometry information is collected and sent (FIG. 18), an empty part may occur in the middle after selecting a bitstream level, and in this case, the bitstream may have to be rearranged.
- unnecessary information can be selectively removed according to the application field as follows.
- 19 illustrates a method of selecting geometry data and attribute data according to embodiments.
- the method/device may select data at the bitstream level as shown in FIG. 21: 1) symmetrical geometry and attribute selection, 2) asymmetrical geometry and attribute selection. choice, 3) or a combination of both methods.
- FIG. 19 it shows a case of transmitting or decoding by selecting only up to LoD1 (LOD 0 +R1, 19000), removing information corresponding to R2 (new part of LOD 2) corresponding to the upper layer, (19010) Transmit and decode.
- a method/device may deliver geometry and attributes asymmetrically. Only the attributes of the upper layer are removed (Attribute R2, 19001), and all of the geometry (from level 0 (root level) to level 7 (leaf level) of the triangular octree structure) is selected for transmission/decoding. can (19011).
- scalable encoding/decoding can be supported when point cloud data is expressed in an octree structure and hierarchically classified for each LOD (or layer).
- the scalability function may include slice level scalability and/or octree level scalability.
- a level of detail (LoD) may be used as a unit for indicating a set of one or a plurality of octree layers.
- it may have the meaning of a bundle of octree layers to configure in slice units.
- An LOD is a unit that divides data in detail by extending the LOD meaning during attribute encoding/decoding, and can be used in a broad sense.
- spatial scalability by the actual octree layer may be provided for each octree layer, but bitstream parsing
- bitstream parsing When scalability is configured in slice units before bitstream parsing, it may be selected in LoD units according to embodiments.
- LOD 0 may be applied from the root level to the 4th level
- LOD 1 may be applied from the root level to the 5th level
- LOD2 may be applied from the root level to the leaf 7 level.
- the provided scalability step is three steps of LoD0, LoD1, and LoD2.
- the scalable steps that can be provided in the decoding step by the octree structure are eight steps from the root to the leaf.
- the transcoder (15040 in FIG. 15 ) of the receiver or transmitter uses 1) only LoD0 for scalable processing. You can select 2) LoD0 and LoD1, or 3) LoD0, LoD1 and LoD2.
- Example 1 When only LoD0 is selected, the maximum octree level is 4, and one scalable layer among octree layers 0 to 4 can be selected in the decoding process. .
- the receiver may consider the node size that can be obtained through the maximum octree depth as a leaf node, and transmit the node size at this time as signaling information. there is.
- Example 2 When LoD0 and LoD1 are selected, layer 5 is added and the maximum octree level is 5, and one scalable layer among octree layers 0 to 5 can be selected in the decoding process. At this time, the receiver may consider the node size that can be obtained through the maximum octree depth as a leaf node, and transmit the node size at this time as signaling information. there is.
- an octree depth, an octree layer, an octree level, and the like refer to units for dividing data in detail.
- Example 3 When LoD0, LoD1, and LoD2 are selected, layers 6 and 7 are added, so the maximum octree level is 7, and one scalable layer among octree layers 0 to 7 (scalable layer) can be selected in the decoding process.
- the receiver may consider the node size that can be obtained through the maximum octree depth as a leaf node, and transmit the node size at this time as signaling information. there is.
- FIG. 20 illustrates a method of constructing a slice including point cloud data according to embodiments.
- a transmission method/device/encoder may be configured by dividing a G-PCC bit stream into a slice structure.
- a data unit for expressing detailed data may be a slice.
- a slice according to embodiments may mean a data unit dividing point cloud data. That is, a slice represents a portion of point cloud data.
- the term slice can be referred to as terms representing a certain part or unit.
- one or a plurality of octree layers may be matched to one slice.
- a transmission method/device may configure a slice 2001-based bitstream by scanning nodes (points) included in an octree in a scan order 2000 direction.
- Some nodes of an octree layer may be included in one slice.
- An octree layer (eg, levels 0 to 4) may constitute one slice 2002 .
- Each slice 2003, 2004, and 2005 may be configured with some data of an octree layer, for example, level 5.
- An octree layer for example, some data of level 6 may constitute each slice.
- An octree layer for example, level 0 to level 3, and some data of level 4 can be configured as one slice.
- An octree layer for example, some data of level 4 and some data of level 5 may be configured as one slice.
- An octree layer for example, some data of level 5 and some data of level 6 may be configured as one slice.
- An octree layer for example, some data of level 6 may be configured as one slice.
- An octree layer for example, data from level 0 to level 4 can be configured as one slice.
- Partial data of each of level 5, level 6, and level 7 of the octree layer may be configured as one slice.
- An encoder and a device corresponding to the encoder may encode point cloud data, generate and transmit a bitstream further including encoded data and parameter information about the point cloud data.
- bitstream when generating a bitstream, the bitstream may be generated based on the bitstream structure (eg, see FIGS. 16 to 20, etc.) according to embodiments. Therefore, a receiving device, a decoder, a corresponding device, etc. according to embodiments may receive and parse a bitstream configured to suit a selective partial data decoding structure, partially decode point cloud data, and efficiently provide it (Fig. 15).
- a method/device for transmitting point cloud data may scalably transmit a bitstream including point cloud data, and a method/device for receiving point cloud data according to embodiments scalably receive a bitstream and can decode.
- Scalable transmission may refer to a case in which only a part of the bitstream is transmitted or decoded rather than decoding the entire bitstream, and the result is low-resolution point cloud data (low resolution point cloud data).
- each octree layer from the root node to the leaf node (FIG. 16 For a bitstream of ), point cloud data must be constructed with only information up to a specific octree layer.
- all octree layers can support scalable transmission, but only scalable transmission for a specific octree layer or lower can make this possible.
- it informs which scalable layer the corresponding slice is included in, thereby determining whether the corresponding slice is necessary/unnecessary at the bitstream stage. can judge In the example of FIG.
- scalable transmission is not supported and one scalable layer is configured, and the following octree layer layer) can be configured to match one-to-one with a scalable layer.
- scalability can be supported for a part corresponding to a leaf node.
- the corresponding layer For layers it can be defined to configure one scalable layer.
- scalable transmission and scalable decoding can be used separately according to the purpose.
- scalable transmission it can be used for the purpose of selecting information up to a specific layer without going through a decoder at the transmitting and receiving end.
- the purpose is to select a specific layer during coding. That is, scalable transmission supports selection of required information without going through a decoder in a compressed state (at the bitstream stage) so that transmission or receiver can determine it.
- encoding and decoding encoding/decoding of only the part required in the encoding and decoding process is supported, such as scalable representation. can be used for
- a layer configuration for scalable transmission and a layer configuration for scalable decoding may be different.
- the lower 3 octree layers including leaf nodes can constitute one layer in terms of scalable transmission, but scalable decoding ) scale for each of the leaf node layer, leaf node layer-1, and leaf node radar-2 when all layer information is included in the perspective.
- Scalable decoding may be possible.
- FIG. 21 illustrates single slice and segmented slice-based geometry tree structures according to embodiments.
- a method/apparatus may configure a slice for transmitting point cloud data as shown in FIG. 21 .
- each slice may contain a sub-bitstream.
- the order of slices may be the same as the order of sub-bitstreams.
- the bitstream is accumulated in breadth-first order of the geometry tree, and each slice can be matched with a group of tree layers (Fig. 21).
- the divided slices may inherit the layering structure of the G-PCC bitstream.
- Slices may not affect previous slices, just as higher layers in a geometry tree do not affect lower layers.
- Segmented slices according to embodiments are efficient in terms of error robustness, effective transmission, and supporting region of interest.
- a divided slice may be more resistant to errors. If a slice contains the entire bitstream of a frame, data loss may affect the entire frame data. Meanwhile, when a bitstream is divided into a plurality of slices, even if some slices are lost, some slices not affected by the loss can be decoded.
- a case in which a plurality of decoders having different capacities may be supported may be considered. If the coded data is in a single slice, the LOD of the coded point cloud can be determined prior to encoding. Accordingly, a plurality of pre-encoded bitstreams having different resolutions of point cloud data may be independently delivered. This can be inefficient in terms of large bandwidth or storage space.
- a single bitstream can support decoders of different levels. Viewed from the decoder side, the receiver can select target layers and deliver the partially selected bitstream to the decoder. Similarly, by using a single PCC bitstream without partitioning the entire bitstream, a partial PCC bitstream can be efficiently generated at the transmitter side.
- a compressed bitstream may be configured to have more than one layer.
- a specific region of interest may have additional layers and high density, and layers may be predicted from lower layers.
- the decoder can increase the resolution of the region of interest upon request. It can be implemented by using scalable structures of G-PCC such as geometry octrees and scalable attribute coding schemes. Based on the current slice structure including the full geometry or attributes, decoders have to access the entire bitstream. This can lead to bandwidth, memory and decoder inefficiencies.
- the decoder slices the bitstream as needed before efficiently parsing the bitstream. can choose
- FIG. 22 illustrates a layer group structure of a geometry coding tree and an aligned layer group structure of an attribute coding tree according to embodiments.
- a method/apparatus may generate a slice layer group using a hierarchical structure of point cloud data as shown in FIG. 22 .
- a method/apparatus may apply segmentation of geometry and attribute bitstreams included in different slices.
- segmentation of geometry and attribute bitstreams included in different slices may be applied.
- a coding tree structure of each slice included in partial tree information and geometry and attribute coding may be used.
- FIG. 22 (a) an example of a geometry tree structure and a proposed slice segment is shown.
- a group represents a group of geometry tree layers.
- text 1 is composed of layers 0 to 4, group 2 includes layer 5, and group 3 includes layer 6 and layer 7.
- group 3 includes layer 6 and layer 7.
- a group can be divided into 3 sub-groups. Parent and child pairs exist in each sub-group.
- Group 3-1 to Group 3-3 are sub-groups of Group 3.
- the tree structure is the same as the geometry tree structure.
- the same octree-slice mapping can be used to build attribute slice segments (Fig. 22(b)).
- Layer group Represents a group of layer structural units occurring in G-PCC coding, such as an octree layer and a LoD layer.
- Sub-group It can be represented as a set of adjacent nodes based on location information for one layer group.
- the lowest layer in the layer group (which may mean a layer closest to the root direction, layer 6 in the case of group 3 in FIG. 22)
- a bundle can be configured based on , and a bundle of adjacent nodes is configured by Morton code order, a bundle of adjacent nodes based on a distance, or a bundle of adjacent nodes according to a coding order. can be configured.
- nodes in a parent-child relationship can be specified to exist in one sub-group.
- a boundary occurs in the middle of the layer, and the entropy continuation enable flag (sps_entropy_continuation_enabled_flag) and entropy determine whether to have continuity at the boundary.
- Continuity with the previous slice can be continuously maintained by notifying whether entropy is continuously used, such as a continuity flag (gsh_entropy_continuation_flag), and notifying a reference slice ID (ref_slice_id).
- FIG. 23 illustrates a layer group and subgroup structure according to embodiments.
- the layer structure-based point cloud data and bitstream shown in FIGS. 21 and 22 may represent a bounding box as shown in FIG. 23 .
- Layer groups 2 and 3 are divided into subgroups 2 (group2-1, group2-2) and 4 subgroups (group3-1, group3-2, group3-3, group3-4) and are included in different slices.
- slices 1, 3, and 6 are selected as layer group 1, subgroup 2-2, and subgroup bounding boxes of 3-3 to cover the ROI region.
- selection and decoding can be performed as each slice segment is received for greater time efficiency.
- the method/apparatus according to the embodiments may express data as a layer (which may be referred to as depth, level, etc.) as a hierarchical tree 23000 when encoding geometry and/or attributes.
- Point cloud data corresponding to each layer (depth/level) may be grouped into a layer group (or group, 2301) as shown in FIGS. 21 and 22 .
- Each layer group may be further divided (segmented) into subgroups 2302 .
- a bitstream may be generated by configuring each subgroup as a slice.
- a receiving device may receive a bitstream, select a specific slice, decode a subgroup included in the slice, and decode a bounding box corresponding to the subgroup.
- the bounding box 2303 corresponding to group 1 can be decoded.
- Group 1 may be data corresponding to the largest area. If the user wants to additionally view the detailed area for group 1, the method/device according to the embodiments selects slice 3 and/or slice 6, and groups 2-2 and/or Alternatively, the bounding boxes (point cloud data) of group 3-3 may be partially accessed hierarchically.
- a method/apparatus determines and selects a slice matching an ROI based on subgroup bounding box information.
- Receivers with different performance can be supported when divided into slices and transmitted as a method of proposing a compressed bitstream for a full coding layer.
- the receiver may select directly or the transcoder may select.
- Information on the case of full decoding when selected by a transcoder eg, total coding layer depth, total number of layer-groups, total subgroups ( number of subgroups, etc.)
- the receiver may need the corresponding information in the decoding process.
- the corresponding information may be directly transmitted or the number of skipped layer groups (num_skipped_layer_groups) and the number of skipped layers (num_skipped_layers) may be delivered as information that can be inferred.
- An example of a method for generating an ROI according to embodiments is as follows.
- ROI setting process for slice selection is as follows (ROI setting for slice selection):
- gbh.layer_group_enabled_flag _sps->layer_group_enabled_flag
- gbh.num_layer_groups_minus1 _sps->num_layer_groups_minus1;
- gbh.num_layers_per_layer_group _sps->num_layers_minus1;
- gbh.num_layers_per_layer_group[i] _sps->num_layers_minus1[i] + 1;
- ROI_size params->roiSize
- ROI_origin[i] int(Width * positionScale + 0.5);
- ROI_origin[i] Width - 1 - ROI_size[i];
- Slice selection supports scalability and spatial random access (// slice selection: scalablity and spatial random access)
- numSkipLayerGroup gbh.num_layer_groups_minus1;
- auto curBboxMin gbh.vec_bboxOrigin[curLayerGroupId][curSubgroupId];
- auto curBboxMax gbh.vec_bboxOrigin[curLayerGroupId][curSubgroupId] + gbh.vec_bboxSize[curLayerGroupId][curSubgroupId];
- PayloadBuffer& buf bufs->at(bufIdx++);
- dep_gbh.geom_parameter_set_id _gps->gps_geom_parameter_set_id;
- dep_gbh.geom_slice_id gbh.geom_slice_id;
- dep_gbh.layer_group_id curLayerGroupId
- dep_gbh.subgroup_id curSubgroupId
- dep_gbh.subgroupBboxOrigin gbh.vec_bboxOrigin[curLayerGroupId][curSubgroupId];
- dep_gbh.subgroupBboxSize gbh.vec_bboxSize[curLayerGroupId][curSubgroupId];
- dep_gbh.ref_layer_group_id gbh.ref_layerGroup[curLayerGroupId][curSubgroupId];
- dep_gbh.ref_subgroup_id gbh.ref_subgroup[curLayerGroupId][curSubgroupId];
- gbh.footer.geom_num_points_minus1 num_points - 1;
- curSubgroupId 0;
- a slice selection method according to the ROI in the decoder is as follows.
- auto curBboxMin _dep_gbh.subgroupBboxOrigin
- auto curBboxMa _dep_gbh.subgroupBboxOrigin + _dep_gbh.subgroupBboxSize;
- auto roiMax _params.roiOrigin + _params.roiSize;
- FIGS. 24 to 26 Examples of resulting images according to the above-described slice selection are shown in FIGS. 24 to 26 .
- FIG. 27 illustrates a slice generation method considering cloud density according to embodiments.
- Method/device for transmitting point cloud data according to embodiments (transmission device 10000 in FIG. 1, point cloud video encoder 10002, transmitter 10003, acquisition-encoding-transmission (20000-20001-20002) in FIG. 2) according to embodiments , encoder of FIG. 4, transmission device of FIG. 12, device of FIG. 14, encoding of FIGS. 15 to 23, encoding of FIGS. 27 to 29, encoding of FIG. 37, encoding of FIGS. 40-42, and transmission method of FIG. 43) are points as shown in FIG. Slices can be created and encoded considering the density of cloud data.
- Method/device for receiving point cloud data are point clouds
- Point cloud data may be decoded based on a slice generated by considering data density.
- the method/device according to the embodiments may use an adaptive subgroup bounding box.
- subgroup split can limit the number of nodes in each sub (sub) group.
- cubic subgroups with fixed subgroup sizes are used within the input sequence. Since division (partitioning) is based on a spatial domain, the range of the number of nodes in each slice is not guaranteed.
- the cubic bounding box can be split into subdivisions as shown in Figure 27. In this example, the original bounding box is split in two by dividing the cubic center on one axis. Dividing across the y and z axes of the three gives the Morton order (order). If the splitting position is not the center position, the sub-partitioned bounding boxes will span Morton order for all three cases.
- the segmentation process may be applied to test content having a large number of points.
- the number of points of data (ulb_unicorn_hires_vox15.ply) is 63,787,119 and the number of subgroups of the last layer group created is 105 (the above-mentioned conditions apply).
- the distribution of points in the point cloud data is not uniform, there may be subgroups whose number of points exceeds 1,100,000 according to TMC13. To adjust the number of points in those subgroups, we can divide the cubic subgroup into several subgroups. Referring to FIG.
- each row (row) represents an example of splitting across the z-axis, y-axis, and x-axis.
- each column (column) indicates the subgroup origin, size, and number of points before/after split.
- the split direction is determined to produce two sub-divided results that have as many similar points as possible.
- the number of points after splitting can be lower than the level limit at which coder friendly slices can be created.
- the number of slices of the last layer group may be changed to 121 after division.
- FIG. 28 An example of division of the cubic bounding box described in FIG. 27 into sub-group boxes is shown in FIG. 28.
- subgroups are divided based on subgroupSize, a parameter that determines the length of a cube. Since the bounding box of the content is divided at a uniform distance in all axes, the cubic subgroup can be regarded as a unit of spatial random access.
- Subgroup generation methods may include 1) a method using a fixed subgroup size and/or 2) a method using both a fixed subgroup size and subgroup division.
- all subgroups may be created based on a fixed subgroup size.
- the following describes layer-group and subgroup based on subgroup origin, subgroup size, point count measurement, and layer-group structure parameter. ) and storing related parameters in SPS (sequence parameter set) and layer-group structure inventory.
- min_k std::round(inputPointCloud.computeBoundingBox().min[k] * params->seqGeomScale);
- gbh_temp.rootNodeSizeLog2[k] ceillog2(std::max(2, bound[k]));
- gbh_temp.rootNodeSizeLog2 gbh_temp.rootNodeSizeLog2.max();
- auto lvlNodeSizeLog2 mkQtBtNodeSizeList(params->gps, params->geom.qtbt, gbh_temp);
- params->subgroupBboxSize_Cubic (params->subgroupBboxSize_Cubic ⁇ (1 ⁇ gbh_temp.rootNodeSizeLog2.max()))
- params->subgroupBboxSize_Cubic 1 ⁇ gbh_temp.rootNodeSizeLog2.max();
- params->subgroupBboxSize_Cubic 1 ⁇ (int)(gbh_temp.rootNodeSizeLog2.max() - 3);
- params->sps.num_layers_minus1[i] params->numLayersInLayerGroup0 - 1;
- params->sps.num_layers_minus1[i] params->num_layers_per_layer_group[i - 1];
- numPerAxis[k] std::ceil(double(1 ⁇ gbh_temp.rootNodeSizeLog2[k]) / initSubgroupBboxSize[k]);
- numMaxSubgroups * numPerAxis[k]
- minInput[k] std::round(minInput[k] * params->seqGeomScale);
- pos[k] std::round(inputPointCloud[i][k] * params->seqGeomScale) - minInput[k];
- subgroupIdx[k] (int)(pos[k] / initSubgroupBboxSize[k]);
- int currentIdx (subgroupIdx[0] * numPerAxis[1] + subgroupIdx[1]) * numPerAxis[2] + subgroupIdx[2];
- curIdx[2] i % numPerAxis[2]
- curIdx[1] (int)(i / numPerAxis[2]) % numPerAxis[1];
- curIdx[0] (int)((int)(i / numPerAxis[2]) / numPerAxis[1]);
- curBboxOrigin[m] initSubgroupBboxSize[m] * curIdx[m];
- params->sps.subgroupBboxOrigin_bits_minus1 numBits(maxXYZ) - 1;
- params->sps.subgroupBboxSize_bits_minus1 numBits(initSubgroupBboxSize.max()) - 1;
- params->sps.subgroupBboxSize_bits_minus1 numBits((int)(1 ⁇ gbh_temp.rootNodeSizeLog2.max())) - 1;
- layerGroupStructureInventory.lgsi_seq_parameter_set_id params->sps.sps_seq_parameter_set_id;
- slice.lgsi_num_layer_groups_minus1 params->sps.num_layer_groups_minus1;
- slice.lgsi_subgroupBboxOrigin_bits_minus1 params->sps.subgroupBboxOrigin_bits_minus1;
- slice.lgsi_subgroupBboxSize_bits_minus1 numBits(1 ⁇ gbh_temp.rootNodeSizeLog2.max()) - 1;
- subentry.lgsi_subgroupBboxOrigin params->subgrpBboxOrigin[idx][i];
- subentry.lgsi_subgroupBboxSize params->subgrpBboxSize[idx][i];
- subentry.lgsi_parent_subgroup_id subentry.lgsi_subgroup_id;
- layerGroupStructureInventory.lgsi_origin_bits_minus1 numBits(_originInCodingCoords.max()) - 1;
- the use of a fixed subgroup size has an advantage in that it is easy to determine the range of a required region in terms of spatial random access, but has a disadvantage in that the number of points cannot be uniformly divided.
- a decoder ie, a level limit or a level limit.
- a method to prevent this a method of dividing subgroups is required. At this time, if the number of subgroups increases, the number of subgroups that can be processed by the decoder may exceed the number of subgroups. Therefore, a method for dividing without significantly increasing the number of subgroups is required. .
- the method/apparatus may divide the cubic subgroup bounding box step by step. That is, when the first constraint condition (the number of points) is satisfied, segmentation is terminated. This reduces the number of additional subgroups compared to splitting in all three directions.
- 29 illustrates a subgroup division process according to embodiments.
- Method/device for transmitting point cloud data according to embodiments (transmission device 10000 in FIG. 1, point cloud video encoder 10002, transmitter 10003, acquisition-encoding-transmission (20000-20001-20002) in FIG. 2) according to embodiments , the encoder of FIG. 4, the transmission device of FIG. 12, the device of FIG. 14, the encoding of FIGS. 15 to 23, the encoding of FIGS. 27 to 29, the encoding of FIGS. 40-42, and the transmission method of FIG. 43) are cubic subs including point cloud data. Groups can be divided as shown in FIG.
- Method/device for receiving point cloud data according to embodiments (receiving device 10004 in FIG. 1, receiver 10005, point cloud video decoder 10006, transmission-decoding-rendering (20002-20003-20004 in FIG. 2)) , decoder of FIGS. 10-11, receiving device of FIG. 13, device of FIG. 14, decoding of FIGS. 15 to 23, decoding of FIGS. 27 to 29, decoding of FIG. 38, decoding of FIGS. 40-42, and receiving method of FIG. 44) are point clouds A cubic subgroup including data may be divided as shown in FIG. 29, or the divided subgroup may be decoded at the transmitting side.
- step 1 to step 3 an example of the proposed subgroup division is shown.
- step 1 to step 3 one of the division directions is selected.
- the first step when the number of points is greater than the level limit, the number of points in the subgroup bounding box is calculated and divided into two subblocks.
- the cost of each split is calculated and compared. To make fewer subgroup splits, the split cost is taken as the difference between the number of points in each sub-division A and B.
- the three costs of x, y, and z-axis division are compared and the direction with the lowest cost is selected as the subgroup division direction.
- the y direction is selected.
- the process ends when the number of points in subdivisions A and B is less than the limit level. On the other hand, if one or two subgroups do not satisfy Constraint 1 (the number of points), the subgroup bounding box is divided in the direction selected in the remainder of the previous step. The same cost function is used for selection criteria. After the second division, the score of each subsection is estimated and compared with the level limit. Depending on the result, the division is terminated in step 2 or additional division is performed.
- a third subgroup division is performed. Since there is only one division direction, subgroup division is performed without selection. The final output of the subgroup division becomes subdivision bounding boxes at each step where the number of points of the bounding box is less than the level limit.
- the following operation is performed by a subgroup divider or processor in charge of subgroup division of an encoder (or decoder).
- the device receives a cubic subgroup containing point cloud data, and as step 1, detects whether the number of points included in the subgroup is greater than a threshold value (level limit). If the number of points is large, a subgroup split is performed. For example, sub-divisions can be performed on three axes, and candidate sub-divisions (or can be referred to as sub-groups) A and B according to the sub-divisions can be generated for each axis. A sub-division having the best score may be selected from among the three candidate sub-divisions.
- a threshold value level limit
- the y-axis sub-division may be performed.
- step 2 when the sub-division having the best score is selected, if the number of points is less than the threshold value, there is no need for further splitting, so the subgroup splitting process is terminated. Otherwise, additional sub-divisions may be performed. For example, since the y-axis sub-division is primarily applied, the y-axis-x-axis subdivision and the y-axis-z-axis subdivision may be considered.
- Scores according to the number of points between the two candidates may be compared, and a sub-division having the best score may be selected.
- step 3 after the secondary sub-division, if the number of points in the sub-group is smaller than the threshold value, since additional division is not required, the sub-division may be terminated. Otherwise, the 3rd subdivision may be performed. For example, the subgroups may be divided into y-axis-z-axis-x-axis. Through steps 1 to 3, sub-divisions (or simply referred to as sub-groups) of subgroups having different sizes may be created.
- the initial subgroup size was used as a point cloud
- it may be created by gradually dividing subgroups by considering the number of points.
- FIG. 30 shows a comparison between a variable subgroup size and a fixed subgroup size according to embodiments.
- Fig. 30 shows the overall segmentation result. As a result, it can be seen that the subgroup division has little effect on the compression efficiency and the number of points in the subgroup is maintained below the level limit.
- mask_base + 1 ⁇ (2 - BestDirection[m]);
- splitBboxOrigin splitBboxSize
- splitBboxOrigin splitBboxSize
- center[m] curBboxOrigin[m] + initSubgroupBboxSize[m] / 2;
- splitIdx + 1 ⁇ (2 - m);
- splitBboxOrigin splitBboxSize
- curBboxOrigin initSubgroupBboxSize, BestDirection, 0, posHigh, maxNumPoint
- splitBboxOrigin splitBboxSize
- curBboxOrigin initSubgroupBboxSize, BestDirection, 0, posHigh, maxNumPoint
- curIdx[2] i % numPerAxis[2]
- curIdx[1] (int)(i / numPerAxis[2]) % numPerAxis[1];
- curIdx[0] (int)((int)(i / numPerAxis[2]) / numPerAxis[1]);
- curBboxOrigin[m] initSubgroupBboxSize[m] * curIdx[m];
- splitSubgroup (splitBboxOrigin, splitBboxSize, subgrpPointCloud[i],
- splitBboxSize push_back(initSubgroupBboxSize);
- params->sps.subgroupBboxOrigin_bits_minus1 numBits(maxXYZ) - 1;
- params->sps.subgroupBboxSize_bits_minus1 numBits(initSubgroupBboxSize.max()) - 1;
- subgroup decoding is dependent on a coding order (coding order)
- subgroup division may not be used. For example, when a node included in a subgroup is imported from a parent, a case in which a parent subgroup includes a plurality of child subgroups may be considered.
- 31 illustrates a subgroup bounding box according to a node scan order according to embodiments.
- Each 4 ⁇ 4 region represents a parent subgroup, and the smallest square represents a node, and the arrow represents the node scan order (Morton order), and the square is indicated by a thick solid line. The line represents the bounding box.
- An example of subgroups whose bounding boxes are constrained in node scan order is illustrated in FIG. 31 .
- Square and rectangular subgroup bounding boxes contain all contiguous nodes shaded.
- a method/device may code subgroup bounding boxes in a predetermined order. It can be seen that all points coded according to the scan order are included in the bounding box.
- 32 illustrates a subgroup bounding box according to a node scan order according to embodiments.
- 32 shows an example of a subgroup in which some shaded nodes are outside the bounding box. This can happen when the size of the boundary is not a power of 2 or when a child (sub)group is created by dividing the parent (parent) subgroup in the Morton order.
- coding may be performed by the following methods.
- a parent subgroup output point included in a bounding box of a child subgroup may be selected. That is, since the parent (higher) subgroup is included in the scan order and boundary, it can be coded according to the scan order of points within the boundary. In this case, necessary points can be selected regardless of the scan order as in the following implementation example.
- auto nodeSizeLog2 lvlNodeSizeLog2[startDepth];
- nodePos[m] node0.pos[m] ⁇ nodeSizeLog2[m];
- rotation can be performed to have the same shape as dividing vertically along the x-axis (eg, 90 degree rotation in the xy plane and 90 degree rotation in the xz plane).
- the decoder may signal that inverse rotation should be performed in the original direction after decoding (eg, -90 degree rotation in the xy plane, -90 degree rotation in the xz plane). .
- the encoder may additionally require a process of reordering points in the rotated subgroup space (eg, Morton code order).
- 33 shows a bitstream including point cloud data and parameters according to embodiments.
- a method/apparatus for transmitting point cloud data may compress point cloud data, generate related parameters, generate a bitstream as shown in FIG. 22, and transmit the bitstream.
- a method/apparatus for receiving point cloud data may receive a bitstream as shown in FIG. 33 and restore point cloud data based on parameter information.
- the method/apparatus according to embodiments may generate signaling information about a radius inter prediction prediction method in predictive geometry according to embodiments.
- information for radius prediction can be defined during inter-screen prediction. Indicates that inter-screen prediction and radius prediction of the predictive geometry node are applied (including) through the sequence parameter set, and the necessary information for this is set according to the implementation method All or part of related information can be delivered to (sequence parameter set).
- each piece of information may be delivered through a geometry parameter set, a slice header (or referred to as a data unit), an SEI message, and the like.
- the application range and application method can be used differently by defining it in a corresponding location or a separate location.
- parameters according to embodiments may be variously called metadata, signaling information, etc.
- parameters according to embodiments may be generated in a process of a transmitter according to embodiments described later, and transmitted to a receiver according to embodiments to be used in a reconfiguration process. It can be.
- parameters according to embodiments may be generated by a metadata processing unit (or metadata generator) of a transmitting device according to embodiments described below and obtained by a metadata parser of a receiving device according to embodiments. .
- An encoded point cloud configuration will be described with reference to FIGS. 22 to 26.
- Tiles or slices are provided so that the point cloud can be divided into areas and processed. When divided into regions, each region may have a different level of importance. By providing that different filters and different filter units can be applied according to their importance, it is possible to provide a method of using a filtering method having high complexity but good quality as a result in an important area. Depending on the processing capacity of the receiver, it is possible to apply different filtering to each area (areas divided into tiles or slices) instead of using complex filtering methods for the entire point cloud, resulting in better image quality in areas that are important to the user. and proper latency on the system can be guaranteed. Accordingly, when the point cloud is divided into tiles, different filters and different filter units may be applied to each tile. When the point cloud is divided into slices, different filters and different filter units may be applied to each slice.
- a bitstream may include SPS, GPS, APS, and TPS.
- APS may be plural.
- the TPS may include tile bounding box related information for a plurality of tiles. For example, location (coordinates) information and size information (width, depth, height) of the bounding box of the tile may be included.
- the bitstream may include geometry information (data) and attribute information (data) in units of slices (data units). Since point cloud data is encoded in units of slices (data units), a bitstream may include a plurality of slices (data units).
- One slice (data unit) may include geometry information (position) of one point and one or more attribute information (color, reflectivity, etc.).
- the 0th slice (data unit) includes geometry data
- the geometry data may include a geometry slice header and geometry slice data.
- the geometry slice header may include information about geometry. For example, including geometry parameter set ID information, geometry tile ID information, geometry slice ID information, origin information of a box (bounding box) including geometry, log scale information of a box, maximum node size information, and number of points information. can do.
- tile parameter set can be defined in conjunction with the coding method, and the tile parameter to support regionally different scalability. It can be defined in a set (tile parameter set).
- the syntax element defined below can be applied to not only the current point cloud data stream but also multiple point cloud data streams, the parameter set of the higher concept ), and so on.
- a bitstream may be selected at a system level by defining a network abstract layer (NAL) unit and delivering related information for selecting a layer such as layer_id.
- NAL network abstract layer
- parameters according to embodiments may be generated in a process of a transmitter according to embodiments described later, and transmitted to a receiver according to embodiments to be used in a reconfiguration process. It can be.
- parameters according to embodiments may be generated by a metadata processing unit (or metadata generator) of a transmitting device according to embodiments described below and obtained by a metadata parser of a receiving device according to embodiments. .
- Fig. 34 shows a set of sequence parameters included in the Fig. 33 bitstream.
- Layer group enable flag (layer_group_enabled_flag): equal to 1 indicates that the geometry bitstream of a frame or tile is included in multiple slices matching the coding layer group or subgroup thereof.
- Layer_group_enabled_flag is equal to 0, it indicates that the geometric bitstream of a frame or tile is included in a single slice.
- Number of layer groups (num_layer_groups_minus1): plus 1 indicates the number of layer groups representing groups of contiguous tree layers in which the layer groups are part of the geometry (geometry) coding tree structure. num_layer_groups_minus1 may range from 0 to the number of coding tree layers.
- the layer group ID (layer_group_id) represents an indicator of a layer group of a frame or tile.
- the range of layer_group_id may be from 0 to num_layer_groups_minus1.
- Number of layers (num_layers_minus1): plus 1 indicates the number of coding layers included in the i-th layer group.
- the total number of layer groups can be derived by adding all (num_layers_minus1[i] + 1) for i equal to 0 to num_layer_groups_minus1.
- Subgroup enable flag (subgroup_enabled_flag): equal to 1, indicating that the current layer group is composed of subgroups that can be included in several slices.
- subgroup_enabled_flag is equal to 0 and indicates that the current layer group is included in a single slice. Subgroups are mutually exclusive and the sum of subgroups is equal to the layer group.
- Subgroup bounding box origin (subgroup_bbox_origin_bits_minus1): plus 1 is the bit length of the syntax element subgroup bounding box origin (subgroup_bbox_origin).
- Subgroup bounding box size (subgroup_bbox_size_bits_minus1): plus 1 is the bit length of the syntax element subgroup bounding box size (subgroup_bbox_size).
- Non-cubic subgroup enable flag (non_cubic_subgroup_enabled_flag): If 1, it may indicate that a non-square subgroup bounding box is used.
- the range of the maximum and minimum values of the subgroup bounding box is subgroup bounding box Max X (subgroup_bbox_max_x), subgroup bounding box Max Y (subgroup_bbox_max_y), subgroup bounding box Max Z (subgroup_bbox_max_z), sub It can be transmitted through the group bounding box minimum X (subgroup_bbox_min_x), subgroup bounding box minimum Y (subgroup_bbox_min_y), and subgroup bounding box minimum Z (subgroup_bbox_min_z) values. That is, the maximum value and the minimum value for each axis of the bounding box on the X-Y-Z axis can be indicated.
- a sequence parameter set may further contain the following elements:
- simple_profile_compatibility_flag 1 specifies that the bitstream conforms to the simple profile.
- simple_profile_compatibility_flag 0 specifies that the bitstream conforms to a profile other than the simple profile.
- dense_profile_compatibility_flag equal to 1 specifies that the bitstream conforms to the Dense profile.
- density_profile_compatibility_flag 0 specifies that the bitstream conforms to a profile other than the Dense profile.
- predictive_profile_compatibility_flag 1 specifies that the bitstream conforms to the predictive profile.
- predictive_profile_compatibility_flag 0 specifies that the bitstream conforms to a profile different from the predictive profile.
- main_profile_compatibility_flag 1 specifies that the bitstream conforms to the default profile.
- main_profile_compatibility_flag 0 specifies that the bitstream conforms to a profile other than the main profile.
- reserved_profile_compatibility_18bits Must be equal to 0 in bitstreams conforming to this version of this document. Another value for reserved_profile_compatibility_18bits is reserved for future use by ISO/IEC. The decoder ignores the value of reserved_profile_compatibility_18bits.
- slice_reordering_constraint_flag equal to 1 indicates that the bitstream is sensitive to reordering and removal of data units.
- slice_reordering_constraint_flag 0 indicates that the bitstream is not sensitive to reordering and removal of data units.
- unique_point_positions_constraint_flag equal to 1 indicates that every output point has a unique position in each point cloud frame that references the current SPS.
- unique_point_positions_constraint_flag 0 indicates that two or more output points may have the same position in any point cloud frame that references the current SPS.
- level_idc Indicates the level to which the bitstream conforms as specified in Annex A.
- the bitstream shall not contain any level_idc values other than those specified in Annex A. Other values of level_idc are reserved for future use by ISO/IEC.
- sps_seq_parameter_set_id Provides an identifier for the SPS so that it can be referenced by other syntax elements. sps_seq_parameter_set_id MUST be 0 in bitstreams conforming to this version of this document. Other values of sps_seq_parameter_set_id are reserved for future use by ISO/IEC.
- frame_ctr_lsb_bits Specifies the length of the frame_ctr_lsb syntax element in bits.
- slice_tag_bits Specifies the length of the slice_tag syntax element in bits.
- seq_origin_bits Specifies the length in bits of the syntax element seq_origin_xyz[ k ].
- seq_origin_xyz[ k ] and seq_origin_log2_scale Specifies the origin of the sequence local coordinate system. Index k is the kth X, Y or Z component of the origin coordinate. If not present, seq_origin_xyz[ k ] and seq_origin_log2_scale values are inferred to be 0.
- the array SeqOrigin is the origin of the sequence's local coordinate system:
- SeqOrigin[k] seq_origin_xyz[k] ⁇ seq_origin_log2_scale
- seq_bounding_box_size_bits The bit length of the syntax element seq_bounding_box_size_minus1_xyz[ k ].
- seq_bounding_box_size_xyz_minus1[ k ]: plus 1 specifies the kth component of the width, height, and depth of the coded volume dimensions, respectively, in the output coordinate system. If not present, the coded volume dimensions are undefined.
- seq_unit_numerator_minus1, seq_unit_denominator_minus1, and seq_unit_in_metres_flag Specifies the length of the output coordinate system X, Y, and Z unit vectors.
- seq_global_scale_factor_log2 Specifies the fixed-point scale factor used to derive the output point position from the position in the sequence local coordinate system.
- seq_global_scale_factor_log2 Used to derive the global scale factor to apply to the location of the point cloud.
- seq_global_scale_refinement_num_bits The bit length of the syntax element seq_global_scale_refinement_factor. If seq_global_scale_refinement_num_bits is equal to 0, no refinement is applied.
- seq_global_scale_refinement_factor Specifies refinement for the global scale value. If not present, seq_global_scale_refinement_factor is inferred to be equal to 0.
- sps_num_attributes Specifies the number of attributes in the coded point cloud. It is a requirement of bitstream conformance that all slices have attribute data units corresponding to all attribute components listed in the SPS.
- attribute_dimension_minus1[ attrId ]: plus 1 specifies the number of elements of the attrId-th attribute.
- attribute_instance_id[ attrId ] Specifies the instance identifier for the attrId-th attribute.
- attribute_bitdepth_minus1[ attrId ]: plus 1 specifies the bit depth of each component of the attrId-th attribute signal(s).
- known_attribute_label_flag[ attrId ] indicates whether the attribute is identified by the value of known_attribute_label [ attrId ] or by the object identifier attribute_label_oid [ attrId ].
- the attribute type identified by known_attribute_label may be specified. If the value of known_attribute_label is not specified, it is reserved for future use by ISO/IEC.
- the attribute type may indicate color, reflectance, opacity, frame index, frame number, material identifier, normal vector, and the like.
- num_attribute_parameters Specifies the number of attribute parameter sets in the bitstream. Attribute parameters that are signaled in the sequence parameter set are applied to all data units of the coded point cloud sequence.
- axis_coding_order Specifies the correspondence between the X, Y and Z output axis labels and the three positional components of every point in the reconstructed point cloud.
- bypass_stream_enabled_flag 1 specifies that the bypass coding mode can be used when reading the bitstream.
- bypass_stream_enabled_flag 0 specifies that the bypass coding mode is not used when reading the bitstream.
- entropy_continuation_enabled_flag 1 indicates that the initial entropy context state of a slice may depend on the final entropy context state of a preceding slice.
- entropy_continuation_enabled_flag 0 specifies that the initial entropy context state of each slice is independent. It is a requirement of bitstream conformance that entropy_continuation_enabled_flag is equal to 0 when slice_reordering_constaint_flag is equal to 0.
- sps_extension_flag 0 specifies that the sps_extension_data_flag syntax element is not present in the SPS syntax structure.
- sps_extension _flag MUST equal 0 in bitstreams conforming to this version of this document.
- the value 1 of sps_extension _flag is reserved for future use by ISO/IEC.
- a decoder MUST ignore all sps_extension_data_flag syntax elements following a value of 1 for sps_extension_flag in the SPS syntax structure.
- sps_extension_data_flag can have any value. Its presence and value do not affect decoder compliance to the profiles specified in Annex A. Decoders conforming to this version of this document MUST ignore all sps_extension_data_flag syntax elements.
- 35 shows a dependent geometry data unit header according to embodiments.
- Fig. 35 shows a geometry data unit header included in the Fig. 33 bitstream.
- Slice ID (dgsh_slice_id): Indicates the ID of the slice for the dependent geometry data unit.
- Layer group ID (layer_group_id): Indicates an indicator of a layer group of a frame or tile.
- the range of layer_group_id may be from 0 to num_layer_groups_minus1.
- the subgroup ID (subgroup_id) represents an indicator of a subgroup of a layer group indicated by the layer group ID (layer_group_id).
- the range of subgroup_id may be between 0 and num_subgroups_minus1 [layer_group_id].
- the subgroup ID (subgroup_id) may indicate the order of slices in the same layer group ID (layer_group_id). If not present, subgroup_id is inferred as 0.
- subgroup_bbox_origin specifies the origin of the subgroup bounding box of the subgroup indicated by subgroup_id of the layer-group indicated by layer_group_id.
- the subgroup bounding box origin (subgroup_bbox_origin) indicates the origin of the subgroup bounding box of the subgroup indicated by subgroup_id of the layer group indicated by layer_group_id.
- the subgroup bounding box size indicates the size of a subgroup bounding box of a subgroup indicated by subgroup_id of a layer group indicated by layer_group_id.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Description
Claims (15)
- 포인트 클라우드 데이터를 인코딩하는 단계; 및상기 포인트 클라우드 데이터를 포함하는 비트스트림을 전송하는 단계; 를 포함하는,포인트 클라우드 데이터 송신 방법.
- 제1항에 있어서,상기 포인트 클라우드 데이터를 인코딩하는 단계는,상기 포인트 클라우드 데이터를 포함하는 바운딩 박스를 서브 그룹 바운딩 박스로 분할하는 단계를 포함하고,상기 서브 그룹 바운딩 박스는 상기 바운딩 박스에 포함된 포인트들의 개수에 기초하여, 상기 바운딩 박스에 관한 축에 기반하여 분할되고,상기 축은 X축, Y축, 및 Z축 중 어느 하나를 포함하고,상기 서브 그룹 바운딩 박스에 포함된 포인트들의 개수는 임계치보다 작고,상기 서브 그룹 바운딩 박스의 개수는 임계치보다 작은,포인트 클라우드 데이터 송신 방법.
- 제2항에 있어서,상기 분할하는 단계는,상기 바운딩 박스에 포함된 포인트들의 개수가 임계치보다 큰 경우, 상기 바운딩 박스를 x축으로 분할하는 서브 디비전, y축으로 분할하는 서브 디비전, z축으로 분할하는 서브 디비전 중에서, 상기 서브 디비전에 관련된 포인트들의 개수에 기초하여, 어느 하나를 선택하고,상기 선택된 서브 디비전에 포함된 포인트들의 개수가 임계치보다 큰 경우, 상기 선택된 서브 디비전을 선택되지 않은 축으로 분할하는 서브 디비전 중에서, 상기 서브 디비전에 관련된 포인트들의 개수에 기초하여, 어느 하나를 선택하고,상기 선택된 서브 디비전에 포함된 포인트들의 개수가 임계치보다 큰 경우, 상기 선택된 서브 디비전을 선택되지 않은 축으로 분할하고,각 서브 디비전의 사이즈는 서로 다른,포인트 클라우드 데이터 송신 방법.
- 제2항에 있어서,상기 포인트 클라우드 데이터를 인코딩하는 단계는,상기 서브 그룹 바운딩 박스에 포함된 포인트들을 몰톤 오더에 따라서 스캔하는,포인트 클라우드 데이터 송신 방법.
- 제2항에 있어서,상기 포인트 클라우드 데이터를 인코딩하는 단계는,상기 서브 그룹 바운딩 박스에 포함된 포인트들을 몰톤 오더에 따라서 스캔하는 도중 포인트가 바운더리를 벗어나는 경우, 상기 포인트의 상위 노드에 포함된 포인트를 스캔하고,상기 서브 그룹 바운딩 박스가 특정 축에 기초하여 생성된 경우, 상기 특정 축에 관련된 평면 상 상기 서브 그룹 바운딩 박스를 90도 회전시키는,포인트 클라우드 데이터 송신 방법.
- 제1항에 있어서,상기 비트스트림은 시퀀스 파라미터 세트, 지오메트리 데이터 유닛 헤더, 및 레이어 그룹 구조 정보를 포함하는,포인트 클라우드 데이터 송신 방법.
- 제1항에 있어서,상기 비트스트림은 서브 그룹 바운딩 박스가 사용되었는지 여부를 나타내는 정보, 및 상기 서브 그룹 바운딩 박스의 위치에 관한 정보를 포함하는,포인트 클라우드 데이터 송신 방법.
- 제1항에 있어서,상기 비트스트림은 서브 그룹 바운딩 박스가 회전하는지 여부를 나타내는 정보, 및 회전 방향을 나타내는 정보를 포함하는,포인트 클라우드 데이터 송신 방법.
- 제1항에 있어서,상기 비트스트림은 서브 그룹 바운딩 박스의 개수, 위치, 크기에 관한 정보를 포함하는,포인트 클라우드 데이터 송신 방법.
- 포인트 클라우드 데이터를 인코딩하는 인코더; 및상기 포인트 클라우드 데이터를 포함하는 비트스트림을 전송하는 트랜스미터; 를 포함하는,포인트 클라우드 데이터 송신 장치.
- 포인트 클라우드 데이터를 포함하는 비트스트림을 수신하는 단계; 및상기 포인트 클라우드 데이터를 디코딩하는 단계; 를 포함하는,포인트 클라우드 데이터 수신 방법.
- 제11항에 있어서,상기 포인트 클라우드 데이터를 디코딩하는 단계는,상기 포인트 클라우드 데이터를 포함하는 바운딩 박스가 분할된 서브 그룹 바운딩 박스를 디코딩하고,상기 서브 그룹 바운딩 박스는 상기 바운딩 박스에 포함된 포인트들의 개수에 기초하여, 상기 바운딩 박스에 관한 축에 기반하여 분할되고,상기 축은 X축, Y축, 및 Z축 중 어느 하나를 포함하고,상기 서브 그룹 바운딩 박스에 포함된 포인트들의 개수는 임계치보다 작고,상기 서브 그룹 바운딩 박스의 개수는 임계치보다 작은,포인트 클라우드 데이터 수신 방법.
- 제12항에 있어서,상기 분할하는 단계는,상기 바운딩 박스에 포함된 포인트들의 개수가 임계치보다 큰 경우, 상기 바운딩 박스를 x축으로 분할하는 서브 디비전, y축으로 분할하는 서브 디비전, z축으로 분할하는 서브 디비전 중에서, 상기 서브 디비전에 관련된 포인트들의 개수에 기초하여, 어느 하나를 선택하고,상기 선택된 서브 디비전에 포함된 포인트들의 개수가 임계치보다 큰 경우, 상기 선택된 서브 디비전을 선택되지 않은 축으로 분할하는 서브 디비전 중에서, 상기 서브 디비전에 관련된 포인트들의 개수에 기초하여, 어느 하나를 선택하고,상기 선택된 서브 디비전에 포함된 포인트들의 개수가 임계치보다 큰 경우, 상기 선택된 서브 디비전을 선택되지 않은 축으로 분할하고,각 서브 디비전의 사이즈는 서로 다른,포인트 클라우드 데이터 수신 방법.
- 제12항에 있어서,상기 포인트 클라우드 데이터를 디코딩하는 단계는,상기 서브 그룹 바운딩 박스에 포함된 포인트들을 몰톤 오더에 따라서 스캔하는,포인트 클라우드 데이터 수신 방법.
- 포인트 클라우드 데이터를 포함하는 비트스트림을 수신하는 리시버; 및상기 포인트 클라우드 데이터를 디코딩하는 디코더; 를 포함하는,포인트 클라우드 데이터 수신 장치.
Priority Applications (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| EP22878957.4A EP4407991A4 (en) | 2021-10-08 | 2022-10-07 | POINT CLOUD DATA TRANSMISSION DEVICE, POINT CLOUD DATA TRANSMISSION METHOD, POINT CLOUD DATA RECEIVING DEVICE AND POINT CLOUD DATA RECEIVING METHOD |
| CN202280066977.6A CN118056405A (zh) | 2021-10-08 | 2022-10-07 | 点云数据发送设备、点云数据发送方法、点云数据接收设备以及点云数据接收方法 |
| KR1020247011367A KR20240056739A (ko) | 2021-10-08 | 2022-10-07 | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 |
| US18/698,199 US12573099B2 (en) | 2021-10-08 | 2022-10-07 | Point cloud data transmission device, point cloud data transmission method, point cloud data reception device, and point cloud data reception method |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| KR10-2021-0134379 | 2021-10-08 | ||
| KR20210134379 | 2021-10-08 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2023059136A1 true WO2023059136A1 (ko) | 2023-04-13 |
Family
ID=85803624
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/KR2022/015149 Ceased WO2023059136A1 (ko) | 2021-10-08 | 2022-10-07 | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US12573099B2 (ko) |
| EP (1) | EP4407991A4 (ko) |
| KR (1) | KR20240056739A (ko) |
| CN (1) | CN118056405A (ko) |
| WO (1) | WO2023059136A1 (ko) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2025067850A1 (en) * | 2023-09-28 | 2025-04-03 | Telefonaktiebolaget Lm Ericsson (Publ) | Systems and methods for point cloud compression |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021261237A1 (ja) * | 2020-06-22 | 2021-12-30 | ソニーグループ株式会社 | 情報処理装置および方法 |
| US20240346707A1 (en) * | 2023-04-13 | 2024-10-17 | Qualcomm Incorporated | Attribute coding and upscaling for point cloud compression |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190355152A1 (en) * | 2017-07-28 | 2019-11-21 | Peking University Shenzen Graduate School | Point cloud attribute compression method based on kd tree and optimized graph transformation |
| US20200020154A1 (en) * | 2017-01-05 | 2020-01-16 | Bricsys Nv | Point Cloud Preprocessing and Rendering |
| US20210004991A1 (en) * | 2019-07-02 | 2021-01-07 | Tencent America LLC | Method and apparatus for point cloud compression |
| US20210183068A1 (en) * | 2019-04-09 | 2021-06-17 | Peking Universtiy Shenzhen Graduate School | Self-Adaptive Point Cloud Stripe Division Method |
| WO2021145573A1 (ko) * | 2020-01-16 | 2021-07-22 | 엘지전자 주식회사 | 포인트 클라우드 데이터 처리 장치 및 방법 |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10430975B2 (en) * | 2016-11-17 | 2019-10-01 | Google Llc | Advanced k-D tree encoding for point clouds by most significant axis selection |
| KR102699245B1 (ko) * | 2019-03-21 | 2024-08-27 | 엘지전자 주식회사 | 포인트 클라우드 데이터 부호화 장치, 포인트 클라우드 데이터 부호화 방법, 포인트 클라우드 데이터 복호화 장치 및 포인트 클라우드 데이터 복호화 방법 |
| CN110418135B (zh) * | 2019-08-05 | 2022-05-27 | 北京大学深圳研究生院 | 一种基于邻居的权重优化的点云帧内预测方法及设备 |
| US11681679B2 (en) * | 2019-09-25 | 2023-06-20 | Atlassian Pty Ltd. | Systems and methods for performing tree-structured dataset operations |
| US11798196B2 (en) * | 2020-01-08 | 2023-10-24 | Apple Inc. | Video-based point cloud compression with predicted patches |
| US12002244B2 (en) * | 2020-04-08 | 2024-06-04 | Qualcomm Incorporated | Global scaling for geometry-based point cloud coding |
| JP7505926B2 (ja) * | 2020-06-18 | 2024-06-25 | Kddi株式会社 | 点群復号装置、点群復号方法及びプログラム |
| WO2022075319A1 (ja) * | 2020-10-07 | 2022-04-14 | パナソニック インテレクチュアル プロパティ コーポレーション オブ アメリカ | 三次元データ符号化方法、三次元データ復号方法、三次元データ符号化装置、及び三次元データ復号装置 |
-
2022
- 2022-10-07 WO PCT/KR2022/015149 patent/WO2023059136A1/ko not_active Ceased
- 2022-10-07 EP EP22878957.4A patent/EP4407991A4/en active Pending
- 2022-10-07 KR KR1020247011367A patent/KR20240056739A/ko active Pending
- 2022-10-07 US US18/698,199 patent/US12573099B2/en active Active
- 2022-10-07 CN CN202280066977.6A patent/CN118056405A/zh active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200020154A1 (en) * | 2017-01-05 | 2020-01-16 | Bricsys Nv | Point Cloud Preprocessing and Rendering |
| US20190355152A1 (en) * | 2017-07-28 | 2019-11-21 | Peking University Shenzen Graduate School | Point cloud attribute compression method based on kd tree and optimized graph transformation |
| US20210183068A1 (en) * | 2019-04-09 | 2021-06-17 | Peking Universtiy Shenzhen Graduate School | Self-Adaptive Point Cloud Stripe Division Method |
| US20210004991A1 (en) * | 2019-07-02 | 2021-01-07 | Tencent America LLC | Method and apparatus for point cloud compression |
| WO2021145573A1 (ko) * | 2020-01-16 | 2021-07-22 | 엘지전자 주식회사 | 포인트 클라우드 데이터 처리 장치 및 방법 |
Non-Patent Citations (1)
| Title |
|---|
| See also references of EP4407991A4 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2025067850A1 (en) * | 2023-09-28 | 2025-04-03 | Telefonaktiebolaget Lm Ericsson (Publ) | Systems and methods for point cloud compression |
Also Published As
| Publication number | Publication date |
|---|---|
| US12573099B2 (en) | 2026-03-10 |
| EP4407991A4 (en) | 2025-01-22 |
| CN118056405A (zh) | 2024-05-17 |
| EP4407991A1 (en) | 2024-07-31 |
| US20240331206A1 (en) | 2024-10-03 |
| KR20240056739A (ko) | 2024-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021066312A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2021049758A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2021002636A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2023287243A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2023287213A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2022015006A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2023136618A1 (ko) | 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 수신 방법 및 포인트 클라우드 데이터 수신 장치 | |
| WO2021029662A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2022098152A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2021141221A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2023059168A1 (ko) | 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 수신 방법 및 포인트 클라우드 데이터 수신 장치 | |
| WO2021132834A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2022149810A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2023136627A1 (ko) | 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 수신 방법 및 포인트 클라우드 데이터 수신 장치 | |
| WO2022225233A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2023059136A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2023054932A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2021246839A1 (ko) | 포인트 클라우드 데이터 처리 디바이스 및 처리 방법 | |
| WO2022075753A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2021080152A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2024014935A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2023075389A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2023075453A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2022092971A1 (ko) | 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 수신 장치 및 포인트 클라우드 데이터 수신 방법 | |
| WO2023059089A1 (ko) | 포인트 클라우드 데이터 송신 방법, 포인트 클라우드 데이터 송신 장치, 포인트 클라우드 데이터 수신 방법 및 포인트 클라우드 데이터 수신 장치 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 22878957 Country of ref document: EP Kind code of ref document: A1 |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 202280066977.6 Country of ref document: CN |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 18698199 Country of ref document: US |
|
| ENP | Entry into the national phase |
Ref document number: 20247011367 Country of ref document: KR Kind code of ref document: A |
|
| ENP | Entry into the national phase |
Ref document number: 2022878957 Country of ref document: EP Effective date: 20240425 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| WWG | Wipo information: grant in national office |
Ref document number: 18698199 Country of ref document: US |