WO2011093688A2 - Procédé et appareil de traitement de signal vidéo - Google Patents

Procédé et appareil de traitement de signal vidéo Download PDF

Info

Publication number
WO2011093688A2
WO2011093688A2 PCT/KR2011/000729 KR2011000729W WO2011093688A2 WO 2011093688 A2 WO2011093688 A2 WO 2011093688A2 KR 2011000729 W KR2011000729 W KR 2011000729W WO 2011093688 A2 WO2011093688 A2 WO 2011093688A2
Authority
WO
WIPO (PCT)
Prior art keywords
motion vector
block
image
motion
picture
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
Application number
PCT/KR2011/000729
Other languages
English (en)
Korean (ko)
Other versions
WO2011093688A3 (fr
Inventor
박승욱
전용준
김정선
최영희
성재원
전병문
박준영
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
LG Electronics Inc
Original Assignee
LG Electronics Inc
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by LG Electronics Inc filed Critical LG Electronics Inc
Publication of WO2011093688A2 publication Critical patent/WO2011093688A2/fr
Publication of WO2011093688A3 publication Critical patent/WO2011093688A3/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

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
    • H04N19/50—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
    • H04N19/513—Processing of motion vectors
    • H04N19/517—Processing of motion vectors by 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
    • H04N19/50—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
    • H04N19/57—Motion estimation characterised by a search window with variable size or shape
    • 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
    • H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H04N19/61—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding

Definitions

  • the present invention relates to a method and apparatus for processing a video signal, and more particularly, to a video signal processing method and apparatus for encoding or decoding a video signal with high efficiency.
  • Compression coding refers to a series of signal processing techniques for transmitting digitized information through a communication line or for storing in a form suitable for a storage medium.
  • the object of compression encoding includes objects such as voice, video, text, and the like.
  • a technique of performing compression encoding on an image is called video image compression.
  • Compression coding on a video signal is performed by removing redundant information in consideration of spatial correlation, temporal correlation, and stochastic correlation.
  • more efficient video signal processing methods and apparatus are required.
  • the present invention was devised to solve the above requirements, and a method of processing a video signal according to the present invention has an object to implement high-efficiency image compression through various motion prediction methods.
  • Another object of the present invention is to provide a motion prediction method considering a blurring phenomenon occurring in an image.
  • Another object of the present invention is to provide a motion vector scaling method in consideration of global and / or local motion occurring in an image.
  • Another object of the present invention is to provide a motion prediction method using an efficient B-picture (or B-slice) direct prediction mode and a B-picture skip mode.
  • Another object of the present invention is to provide an encoder device and a decoder device for implementing the motion prediction method.
  • a video signal processing method includes acquiring scaled motion vector information of a block to be decoded, acquiring a global motion vector of a neighboring picture, and global motion of the neighboring picture Restoring the scaled motion vector to the original motion vector using a vector, and decoding the corresponding block using the restored original motion vector.
  • the video signal processing method includes filtering the reference picture and the current coded picture in consideration of a blurring phenomenon occurring in the picture.
  • the video signal processing method includes the step of considering the partition result of the neighboring block in the block partition (block partition). It also includes the step of considering the edge (edge) of the peripheral block.
  • the video signal processing method includes a block partition step using a curve.
  • the video signal processing method includes setting a variable search range proportional to a time distance and applying the same directly to a prediction mode.
  • efficient image encoding and decoding are possible. For example, it is possible to prevent the distortion due to the blurring phenomenon, it is possible to prevent the distortion due to the acceleration movement occurring in the entire image or some objects.
  • by setting the search range having a variable size in proportion to the time distance it is possible to prevent distortion in the direct prediction mode.
  • FIG. 1 schematically shows an encoder device applicable to the present invention.
  • FIG. 2 schematically shows a decoder device applicable to the present invention.
  • 3 to 4 illustrate a motion prediction method in consideration of a blurring phenomenon according to an embodiment of the present invention.
  • 5 to 8 illustrate a motion prediction method in consideration of the degree of motion of an image according to another embodiment of the present invention.
  • 9 to 13 illustrate a block partitioning method considering a partition and an edge of a neighboring block according to another embodiment of the present invention.
  • 21 to 22 illustrate a motion prediction method in consideration of applying a direct prediction mode.
  • Coding may be interpreted as encoding or decoding in some cases, and information is a term including all values, parameters, coefficients, elements, and the like. As the meaning can be interpreted differently according to the present invention is not limited thereto.
  • the unit is used to mean a basic unit of image processing or a specific position of an image, and in some cases, the unit may be used interchangeably with terms such as a block or an area.
  • the encoding apparatus 100 of the present invention is largely composed of a transformer 110, a quantizer 115, an inverse quantizer 120, an inverse transformer 125, a filter 130, and a predictor ( 150 and the entropy coding unit 160.
  • the converter 110 obtains a transform coefficient value by converting a pixel value of an input video signal or a residual signal between the video signal and a motion predicted / compensated image.
  • a transform coefficient value For example, Discrete Cosine Transform (DCT) or Wavelet Transform (DCT) may be used.
  • the discrete cosine transform divides the input image signal into blocks having a predetermined size and performs the conversion.
  • the coding efficiency may vary depending on the distribution and the characteristics of the values in the transform domain.
  • the quantization unit 115 quantizes the transform coefficient value output from the transform unit 110.
  • the inverse quantization unit 120 inverse quantizes the transform coefficient value, and the inverse transform unit 125 restores the original pixel value by using the inverse quantized transform coefficient value.
  • the filtering unit 130 performs a filtering operation for improving the quality of the reconstructed image.
  • a deblocking filter and / or an adaptive loop filter may be included.
  • the filtered image is stored in the frame storage unit 156 for output or use as a reference image.
  • a method of predicting an image by using an already coded region and adding a residual value between the original image and the predicted image to a reconstructed image is used instead of coding the image signal as it is.
  • the intra predictor 152 performs intra prediction within the current image
  • the inter predictor 154 predicts the current image using at least one reference image stored in the frame storage 156.
  • the intra predictor 152 performs the intra prediction from the reconstructed regions in the current image and transmits the intra encoding information to the entropy coding unit 160.
  • the inter predictor 154 may further include a motion compensator 162 and a motion estimator 164.
  • the motion estimator 164 acquires a motion vector value of a region to be currently encoded by using reference images stored in the frame storage.
  • the motion estimator 164 transmits the position information (reference frame, motion vector, etc.) of the reference region to the entropy coding unit 160 so that the motion estimation unit 164 may be included in the bitstream.
  • the motion compensator 162 performs inter-screen motion compensation by using the motion vector value transmitted from the motion estimator 164.
  • the entropy coding unit 160 entropy codes the quantized transform coefficients, inter picture encoding information, intra picture encoding information, and reference region information input from the inter prediction unit 154 to generate a video signal bitstream.
  • the entropy coding unit 160 may use a variable length coding (VLC) method, arithmetic coding, or the like.
  • VLC variable length coding
  • the variable length coding (VLC) scheme converts input symbols into consecutive codewords, which may have a variable length. For example, symbols that occur frequently are represented by short codewords and symbols that do not occur frequently by long codewords.
  • a context-based adaptive variable length coding (CAVLC) method may be used as a variable length coding method.
  • Arithmetic coding converts consecutive data symbols into a single prime number, which arithmetic coding can obtain the optimal fractional bits needed to represent each symbol.
  • Context-based Adaptive Binary Arithmetic Code (CABAC) may be used as arithmetic coding.
  • the decoding apparatus 200 of the present invention largely includes an entropy decoding unit 210, an inverse quantization unit 220, an inverse transform unit 225, a filtering unit 230, and a prediction unit 250.
  • the entropy decoding unit 210 entropy decodes the video signal bitstream and extracts a coding type, transform coefficients for each region, a motion vector, and the like.
  • the inverse quantization unit 220 inverse quantizes the entropy decoded transform coefficient, and the inverse transform unit 225 restores the original pixel value by using the inverse quantized transform coefficient.
  • the predictor 250 includes an intra predictor 252 and an inter predictor 254, and includes an encoding type decoded by the entropy decoder 210 described above, transform coefficients for each region, a motion vector, and the like. The predicted image is reconstructed using the information of.
  • the intra prediction unit 252 generates the intra prediction image from the decoded sample in the current image.
  • the inter prediction unit 254 generates a prediction image using the reference image stored in the frame storage unit 256.
  • the reconstructed video frame is generated by adding the pixel value output from the inverse transformer 225 to the predicted image output from the intra predictor 252 or the inter predictor 254.
  • the filtering unit 230 performs filtering on the restored video frame to improve the image quality. This may include a deblocking filter for reducing block distortion and an adaptive loop filter for removing distortion of the entire image.
  • the filtered image is output as the final image and stored in the frame storage unit 256 for use as a reference image for the next frame.
  • a motion prediction method considering the blurring phenomenon with reference to FIGS. 3 to 4, global and / or local (see FIG. 5 to FIG. 8).
  • a motion vector scaling method considering local motion a block partition method with reference to FIGS. 9 to 17, a motion vector signaling method with reference to FIGS. 18 to 20, and FIGS. 21 to 22
  • a detailed description will be given of a prediction method using an efficient B-picture (or B-slice) direct prediction mode and a B-picture skip mode.
  • FIG. 3 illustrates a motion prediction method according to an embodiment of the present invention.
  • FIG. 3 illustrates a motion prediction method in consideration of a blurring phenomenon.
  • FIG. 3 (a) illustrates a reference image 301 and a reference block 301a in the image, for example, showing a clear image in which blurring does not occur.
  • a blurring phenomenon occurs in the reference image, it is possible to apply the motion prediction method according to the present invention, which will be described later.
  • FIGS. 3B, 3C, and 3D show current images 311, 3112, and 313 as current encoding targets, and coding blocks 311a, 312a, and 313a in the corresponding images, respectively.
  • FIG. 3 (a) illustrates an example of an image in which Gaussian blurred occurs due to out of focus in an image capturing means (for example, a 'camera').
  • 3 (b) shows an example of an image in which motion blur occurs due to camera shake or the like during image capturing
  • FIG. 3 (c) shows zoom blurring according to image magnification during image capturing.
  • Each of the examples of the image where) is generated is shown.
  • 3 (b), (c) and (d) are merely examples of various examples in which blurring occurs for convenience of description, and the present invention is limited to only the above type of blurring. It is not.
  • the above-described blurring phenomenon may occur due to various causes during image capturing.
  • the conventional video encoding method does not consider such a blurring phenomenon, and deterioration occurs in the reconstructed picture quality as the video encoding is performed, or a lot of bits of information are required to heal the problem, thereby reducing the coding efficiency. Will occur.
  • the prediction reference blocks of the current coding blocks 311a, 312a, and 313a may be estimated as block 301a in FIG. 3 (a) and a motion vector may be obtained.
  • the block 301a in Fig. 3A is the predicted reference block of the current coding blocks 311a, 312a, and 313a. Even if it is not estimated, or even if it is estimated, there is a problem of inevitably having many prediction errors.
  • the present invention proposes a motion prediction method reflecting the blurring phenomenon in motion prediction in order to prevent the problems caused by the blurring phenomenon as described above.
  • a new reference image reflecting the blurring phenomenon may be generated and used for motion prediction.
  • the motion prediction method reflecting the blurring phenomenon will be described in detail with reference to FIG. 4.
  • FIG. 4 is a flowchart illustrating a motion estimation method reflecting a blurring phenomenon according to an embodiment of the present invention.
  • a comparison between the reference image 301 and the current image determines whether blurring occurs and the type of blurring (S101).
  • S101 for example, it may be determined whether a blurring phenomenon occurs in a specific area or the entire area of the image and the type of blur using a conventionally known local power spectrum method.
  • step S101 it is checked whether the blurring phenomenon identified through the step S101 occurs in the reference image or the current image (S102).
  • the step S102 can be confirmed through relative comparison. That is, for example, if the reference picture is relatively blurred than the current picture, it is determined that there is a blurring phenomenon in the reference picture, and in the opposite case, it is determined that there is blurring in the current picture.
  • step S102 If it is determined in step S102 that there is a blurring phenomenon in the current image, the reference image is blurred according to the blurring type and the blurring degree of the current image and a new reference image reflecting the blurring phenomenon is generated ( S103). On the other hand, if it is determined in step S102 that there is a blurring phenomenon in the reference image, the reference image is de-blurred to generate a new reference image from which the blurring phenomenon is removed (S104). However, in steps S103 and S104, it is preferable to match the reference image with the same or at least similarly to the image state of the current image for motion prediction, so that the level (or level) or deblur blurs the reference image (S103). The level (or level) of ring S104 can be determined by the screen state of the current video.
  • step S104 the blurring phenomenon of the reference image is completely removed. De-blurring may be performed. Also, for example, if there is no blurring in the reference picture (eg, No blurring) and a certain level of blurring occurs in the current picture, the blurring of the reference picture in the current picture level occurs in step S103. A new reference picture is generated.
  • the blurring of the current image level is applied to the reference image through step S103.
  • a new reference image generated is generated.
  • the degree of blurring of the reference picture in the current picture level is performed through step S104. Debbling is performed to reach a new reference picture.
  • the blurring and deblurring performed in step S103 or step S104 may be performed through filtering using filter coefficients.
  • filtering may be applied only to a part of the reference picture instead of the entire area.
  • step S103 or step S104 motion estimation and motion compensation on a block in the current image are performed using the newly generated reference image (eg, the filtered reference image) (S105).
  • the above-described blurring related information eg, blurring type, blurring filter coefficient, blurring parameter, etc.
  • the decoder it is possible to reconstruct a reference image by using the blurring related information used in the encoding process and to use the image for subsequent motion compensation.
  • the new reference image generated in steps S103 and S104 is stored in the frame storage unit 156 in place of the existing reference image, or stored in the frame storage unit 156 as a reference image separate from the existing reference image. This is all possible and it is possible to adapt adaptively by the coding design designer.
  • FIG. 5 to 8 illustrate a motion estimation method according to another embodiment of the present invention, and more particularly, to a motion vector scaling method in consideration of global and / or local motion of an image.
  • the motion vector scaling method proposed by the present invention is particularly effective when there is acceleration movement movement in the global image or in some objects, but is not limited thereto.
  • recent image coding methods introduce cases in which a plurality of reference images are used for motion prediction. That is, for example, it is possible to select an optimal reference block by utilizing different reference images for each coding block in the same picture (or picture), thereby enabling more efficient image compression.
  • determining the motion vector of the current coding block a motion of fitting the motion block of the neighboring block to the current coding block to obtain a motion vector prediction (mvp) using the motion vector of the block on which the neighboring coding is completed.
  • Scaling methods are also introduced.
  • the present invention aims to propose a method which considers global and / or local motion of an image, in particular, in applying the motion vector scaling. This suggests a more efficient method than simply considering only temporal distance, and may be particularly effective when there is accelerated movement movement in the global image or in some objects.
  • the embodiments considering the global motion and the embodiments considering the local motion will be described separately.
  • POC Picture Order Count
  • FIG. 5 illustrates an example in which a motion vector scaling method of the present invention is presented in syntax.
  • a motion vector scaling method may include a global motion vector (gmv) representing a global motion of an image (picture) corresponding to each POC. It is characterized by utilizing.
  • the global motion vector gmv may be, for example, an average or a median obtained by averaging motion vector values of each block in the corresponding image.
  • the syntax and motion vector scaling of FIG. 6 are performed by using the global motion vector gmv obtained for each image (picture).
  • motion vector scaling is performed in a list of reference pictures having a POC value smaller than the POC value of a video to be encoded (named 'reference list_0' for briefly, 'List0'). Show the case.
  • the motion vector scaling is performed in a list of reference pictures having a POC value larger than the POC value of the current video encoding object (hereinafter, referred to as 'reference list_1' or 'List1'). Shows the case.
  • Step S601 is applied to the aforementioned reference images in 'List0' and includes steps S6011 and S6012.
  • the motion vector mvX ' that scales the first motion vector mvX and 502 to the base motion vector mvBase and 511 is the first motion vector mvX and 502 and the base motion vector mvBase and 511.
  • mvX ' mvBase-gmv (14)-gmv (12). That is, by applying the above-described global motion vector existing between the first motion vector mvX and 502 and the base motion vector mvBase, 511 to motion vector scaling, it is possible to correspond to the degree of motion more accurately.
  • the motion vector mvX ' that scales the first motion vector mvX and 503 to the base motion vector mvBase and 511 may include the first motion vector mvX and 503 and the base motion vector mvBase.
  • mvX ' mvBase + gmv (18) + gmv (16). That is, by applying the above-described global motion vector existing between the first motion vectors mvX and 503 and the base motion vectors mvBase and 511 to the motion vector scaling, it is possible to more accurately correspond to the degree of motion.
  • Step S602 is applied to the above-mentioned reference pictures in 'List1' and includes steps S6021 and S6022.
  • the motion vector mvX ' that scales the first motion vector mvX and 504 to the base motion vector mvBase and 512 includes the first motion vector mvX and 504 and the base motion vector mvBase. , 512) may be determined by the following equation in consideration of the degree of movement.
  • mvX ' mvBase + gmv (24) + gmv (26).
  • the motion vector mvX ' that scales the first motion vector mvX and 505 to the base motion vector mvBase and 512 may include the first motion vector mvX and 505 and the base motion vector (mvX').
  • the decoding of the encoded video signal as described above generally proceeds in the reverse of the encoding process. For example, acquiring scaled motion vector information of a block to be decoded, acquiring a global motion vector of a neighboring picture, and using the global motion vector of the neighboring picture, the original motion is converted into the scaled motion vector. Reconstructing a vector, and decoding the block to be decoded by using the reconstructed original motion vector.
  • FIG. 7 illustrates a method of scaling a motion vector in consideration of local motion of some objects in an image.
  • the present embodiment can be effectively applied to the acceleration motion of the object, but does not have a limitation that is applied only during the acceleration motion.
  • a method more efficient than the method using the global motion vector (gmv) considering the global movement of the entire image is proposed as in the above-described embodiment of FIG. That is, in order to consider local motion, a method of tracking motion of blocks around a corresponding object and applying it to motion vector scaling is proposed.
  • FIG. 7 (a) illustrates an example of checking the degree of movement of a specific object through a motion vector value of a neighboring block.
  • a reference object 0 hereinafter referred to as 'ref 0'
  • a current image hereinafter, referred to as 'ref 0'
  • 'ref 1' a movement of a specific object in reference image 1
  • FIG. 2 there are shown cases where 7021 and 7011 are moved.
  • the movement of the objects 7031, 7021, 7011 illustrated in FIG. 7A may be understood as a kind of acceleration motion.
  • the current coding block 713 (current block) in the current video is 'ref 0'
  • the neighboring left block 712 (coded block) is 'ref 0'
  • the neighboring upper block 711 (coding) is used.
  • This can be expressed as a temporal distance, eg, it is necessary to scale d2) to 'ref 0'. That is, assuming that the blocks 711, 712, 713 are all present in the same object 7011, the upper block 711 utilizes the temporal distance value (ie d2 ') is proposed using the left block 712. This will be described in detail below.
  • the scaled motion vector distance value d2 ' means that the scaled motion vector distance value d2' may be scaled at the same ratio as the proportional value of the motion vector distance values d1 and d1 'of another block in the same object.
  • the motion vector difference (mvd: motion vector difference) can be reduced. Efficient coding becomes possible.
  • the utilization of the motion vector distance value (d: distance) is merely an example of the present embodiment, for example, the distance used in applying the motion vector scaling by the same method as the above embodiment. It is obvious that the sum (or difference) of actual motion vectors can be obtained instead of the value (d: distance).
  • a method of predicting d2' by comparing the value obtained by (d2-d1) with the size of d1 ' is also applicable.
  • FIG. 8 illustrates an embodiment of a method of determining an actual motion vector value from the motion vector distance value d identified through FIGS. 7A to 7C.
  • a macro block (MB: Macro Block, 810, 811, 812, overlapped with the corresponding location 801) is used.
  • the most overlapped portion of the 813 determines the macro block.
  • the macro block '810' is the macro block in which the largest portion overlaps.
  • the partition part (eg, 810b) occupying the most overlap part is determined, and the corresponding It is possible to determine the motion vector value applied to the partition 810b as the motion vector value corresponding to the distance value d.
  • FIG. 9 to 17 illustrate a block partitioning method for determining a prediction unit (PU) according to another embodiment of the present invention.
  • FIG. 9 illustrates, for example, a structure of a coding unit (CU) and a corresponding prediction unit (PU) in a video encoding scheme applicable to an embodiment of the present invention. It is.
  • the size of the coding unit CU may be determined as a macro block (e.g., 2N ⁇ 2N) of any size.
  • a macro block e.g., 2N ⁇ 2N
  • FIG. 9 exemplarily illustrates a case in which a total of seven coding units 901, 903, 904, 9021, 9022, 90223, and 9024 are generated.
  • the prediction unit PU for applying motion prediction in the coding unit CU may be determined by various methods. This makes it possible to set the friction unit in various ways in consideration of the characteristics of the input image, for example, an edge, a movement degree, an object or a background image, and thus a high quality and high efficiency image compression effect. It can be increased.
  • dividing the coding unit block to generate the prediction unit (PU) is called a block partition. However, this is only meant as a term, it will be obvious that it can be used in other terms in actual application.
  • FIG. 9 illustrates three types of block partition types.
  • the first partition type is referred to as a symmetric type (911, symmetric type), and is a case of partitioning the coding unit symmetrically into the same size prediction unit (PU) (9111, 9112, 9113, 9114).
  • the second partition type is referred to as an asymmetric type (912, asymmetric type), and is a case in which the coding unit partitions asymmetrically into different sized prediction units (PUs) (9121, 9122, 9123, 9124).
  • the last third partition type is called a geometric type 913, which divides the coding unit into slanted diagonal lines 913a rather than vertical or horizontal straight lines to generate two prediction units 9131 and 9132. That's the way.
  • a method of generating a geometric type 913 and a new geometric type as the third partition type will be described later in detail with reference to FIGS. 14 to 17.
  • 10 to 13 show an example of a method of performing the block partition according to another embodiment of the present invention.
  • 10 (a) to 10 (d) illustrate a method of determining a block partition, wherein a partition of the current block is determined along a partition line of adjacent neighboring blocks. That is, since there is a high probability that similar images exist between neighboring blocks in the same image, determining partition lines of the current block in consideration of partition lines of neighboring neighboring blocks enables efficient image compression.
  • FIG. 10A illustrates an example of a block partition of a symmetric type, in accordance with partition lines 1001a and 1002a of neighboring blocks 1001 and 1002, and partition lines 1003a and 1003b of a current block 1003.
  • partition line 1001a of the current block 1003 is determined by extending the peripheral top block 1001 and the partition line 1001a, and by extending the peripheral left block 1002 and the partition line 1002a.
  • the partition line '1003b' of the current block 1003 may be determined. Therefore, a total of three friction units (PUs) are generated in the current block 1003.
  • PUs friction units
  • FIG. 10B illustrates an example of block partitions of an asymmetric type, in accordance with partition lines 1010a, 1010b, and 1010c of a neighboring block 1010, and partition lines 10201a and 1020b of a current block 1020.
  • the partition line '1020a' of the current block 1020 is determined by extending the peripheral top block 1010 partition line 1010a, and another partition line 1010c is extended to extend the current block 1003.
  • Partition line '1020b' can be determined. Accordingly, a total of three friction units (PUs) are generated in the current block 1020.
  • PUs friction units
  • 10 (c) and 10 (d) are examples of block partitions of an asymmetric type, and according to partition lines 1030a and 1031a of neighboring blocks 1030 and 1031, respectively, of the current blocks 1040 and 1041.
  • the case where the partition lines 1040a and 1041a are respectively determined is shown. That is, for example, in FIG. 10C, the partition line 1030a of the peripheral upper block 1030 is extended to determine the partition line 1040a of the current block 1040.
  • the partition line 1041a of the neighboring top block 1031 is extended to determine the partition line 1041a of the current block 1041.
  • 11 to 13 illustrate an example of a method of adjusting or adjusting a partition line determined by a block partition.
  • the partition line 1101a of the current block 1101 is adjusted to the new partition line 1101b according to the existence of the edge 1102a in the neighboring block 1102.
  • the adjusted partition line '1101b' may be determined by extending the boundary region 1102a line of the neighboring block.
  • FIG. 11A illustrates a case in which the partition line is adjusted in the horizontal direction, it will be apparent that the same may be applied to the vertical direction.
  • the existence of an edge in the neighboring upper block 1102 and the direction of the edge are determined on a pixel-by-pixel basis. It becomes possible.
  • the 'Sobel operator' is generally known as an arithmetic tool used for edge detection.
  • 11 (b) and 11 (c) illustrate, for example, that an adjustment range of the partition line can be limited to N pixels or N / 2 pixels. Therefore, although not shown, it is also possible to limit the partition line adjustment range to N / 4 pixels or N / 8 pixels.
  • the partition line 1201a of the current block 1201 is adjusted to the new partition line 1201b according to the existence of the edge 1202a in the neighboring block 1202.
  • the adjusted partition line '1201b' may be determined by extending the boundary area 1202a line of the neighboring block.
  • FIG. 12A illustrates a case in which the adjustment of the partition line adjusts only one of a start point and an end point of the line. In connection with FIG. 12A, only the end point of the partition line is adjusted in the horizontal direction, for example. However, it is obvious that the same applies to the adjustment of the vertical direction and / or the start point. will be.
  • an adjustment range of the partition line can be limited to N pixels or N / 2 pixels. Therefore, although not shown, it is also possible to limit the partition line adjustment range to N / 4 pixels or N / 8 pixels.
  • the partition line 1301a of the current block 1301 is adjusted to a new partition line 1301b according to the existence of the edge 1302a in the neighboring block 1302.
  • the adjusted partition line 1301b may be determined by extending a line of the boundary area 1302a of the neighboring block.
  • FIG. 13A shows a case where the adjustment of the partition line is applied to both the start point and the end point of the line.
  • the case where the start point and the end point of the partition line are adjusted in the horizontal direction is illustrated as an example. However, it will be apparent that the same applies to the adjustment in the vertical direction. .
  • the adjustment of the partition line can be performed both at the encoder device or the decoder device, but can be preferably designed to be performed at the decoder device.
  • the result of performing the block partition of FIGS. 11 to 13 may be generated and transmitted as bitstream information.
  • the bitstream information may include information (e.g., '1-bit flag') indicating whether the partition line adjustment is applied.
  • the partitioning method of the neighboring block only the information to use the partition of the current block may be transmitted, and information about the current block may not be transmitted.
  • the geometric partitioning exists as a type of block partition.
  • the geometric partitioning is a method of increasing coding efficiency when the contour of an object is included in a block.
  • the contour of the object has an arbitrary shape rather than a straight line, it may be more efficient to express a block as a curved line rather than a straight line when dividing a block into two regions.
  • the method proposed in the present invention is to provide a method for utilizing a curve as well as a straight line in the aforementioned geometric partitioning.
  • the block partition in the current block 1401 may be divided into a straight line (e.g., 1401a) and a curved line (e.g., 1401b).
  • the block partition may be represented by the slopes ⁇ and 1412 and the distance ⁇ 1413 around the center point 1411 of the block.
  • the curvature parameters ⁇ 1414 may be further considered in order to implement a curved partition.
  • FIG. 15 is a diagram illustrating a straight line 1401a dividing a block in FIG. 14 by rotating 90- [theta] to be horizontal for convenience of description.
  • the length of the straight line 1401a is assumed to be '2L'.
  • ⁇ curvature parameter
  • the decoder takes three parameters, calculates a curve equation, and converts the coordinates p (x, y) of each pixel in the block 1401 to p '(x', y ') using the following equation.
  • P ' Rp + T, where R is cos (90- ⁇ ), -sin (90- ⁇ ), sin (90- ⁇ ), cos (90- ⁇ ) Is a 2x2 matrix, and T may be a 2x1 matrix having tx and ty as elements.
  • tx and ty denote coordinates of the center point G of the straight line 1401a in the block. If the value calculated by substituting the calculated (x ', y') into y'-cx ' ⁇ 2-d is less than' 0 ', p (x, y) is determined based on the curve 1401b. ) Is divided into two areas and belongs to the area including the center point 1411, and if it is greater than '0', it belongs to the opposite area.
  • the parameter ⁇ which determines the curvature, represents the distance from the center point (G) of the straight line to the curve measured in the vertical direction. This value can be expressed as an integer, and the range of the value can be adjusted according to the size of the block. Will be. For example, if the block is 16x16, the range of ⁇ is set to have a size of about -4 to +4, and if the block is 32x32, the range of ⁇ is set to have a size of about -8 to +8. It is possible to do
  • FIG. 16 illustrates an example of a table previously stored in an encoder and a decoder device to perform curved partitioning in the aforementioned geometric partitioning. That is, for example, it is possible to easily implement a block partition by storing the parameter values necessary to implement the curved partition for each block size (e.g., 16x16, 32x32, etc.) in advance and storing them in a table.
  • block size e.g. 16x16, 32x32, etc.
  • FIG. 17 illustrates a method of explicitly expressing a complex boundary area in the block 1701 resulting from performing the curved partitioning in the aforementioned geometric partitioning using a chain code 1702. It is shown.
  • FIG. 17 for example, when there are 13 pixels e0 to e12 in contact with a boundary area within a block when the size of the current block is 8x8, the rest of the rest is assumed based on the position of e0.
  • the positions e1 to e12 can be simply expressed by chain codes that are sequentially defined.
  • the above-described curved partition may be implemented as a partition composed of a plurality of straight lines instead of a perfect curve.
  • the parameter ⁇ 1 is set at the center of the straight line and the parameters ⁇ 2 and ⁇ 3 are set at quarter lengths at each end of the straight line, a total of five points are generated (not shown), so it is changed to a curve or , Or you can create a straight line connecting five points.
  • the present invention is to provide a variety of signaling (signaling) method for transmitting the motion vector is included in the bitstream in the motion prediction used up to 1/8 pixel units.
  • FIG. 18 shows a first method for motion vector signaling according to the present invention.
  • FIG. 19 illustrates a second method for motion vector signaling according to the present invention.
  • the motion vector adjustment information for the 1/8 pixel unit may include information transmitting whether 1/8 pixel is applied to each of the X and Y axes.
  • the first method is determined to proceed in units of 1/8 pixels, the motion vector itself is generated and transmitted as such, and the second method is in units of 1/8 pixels.
  • Motion vector refinement information is transmitted for the X and Y axes, respectively. That is, only one bit is transmitted according to the 'MV resolution flag' (FIG. 18), or three bits including two bits of motion vector refinement information are transmitted.
  • FIG. 20 illustrates a third method for motion vector signaling according to the present invention.
  • FIG. 20 illustrates a case in which only motion vector refinement information for 1/8 pixel units is directly signaled without the 'MV resolution flag'.
  • the motion vector adjustment information for the 1/8 pixel unit may include information transmitting whether 1/8 pixel is applied to each of the X and Y axes. In other words, only the 2-bit motion vector refinement information is transmitted.
  • 21 to 22 illustrate a motion prediction method using an efficient B-picture (or B-slice) direct prediction mode and a B-picture skip mode according to another embodiment of the present invention. It is shown.
  • FIG. 21 is a diagram to describe a method applied to a direct prediction mode and a skip mode according to the present invention.
  • the application of a B direct mode and a B skip mode is allowed for a B-picture image.
  • the B direct mode and the B skip mode make it possible to obtain motion vector information without performing full bidirectional (or bi-prediction) prediction in a block to be encoded in a B-picture. do.
  • a size of a search range for predicting a reference block is set differently in proportion to the mutual distance. I need a way to do it.
  • the search range 2102 in 'ref_0 in List (0)' is 16 pixels in either the horizontal or vertical direction
  • the search range (2103, search range) in 'ref_2 in List (1)' is about 3 times farther away, so that it is equally tripled in either horizontal or vertical direction. It characterized in that it is set to.
  • FIG. 22 illustrates a point symmetry method and an asymmetry method for solving a problem in point symmetry search as described above with respect to the above-described B direct mode and B skip mode. Suggest a mix.
  • the search range is set in proportion to the distance to improve the problem in the point symmetry method as shown in FIG. 21 described above, it will be possible to apply the point symmetry method irrespective of the distance.
  • the search range 2103 is set proportional to the distance which improves the problem in the point symmetry method as shown in FIG. 21 described above, it will be possible to apply the point symmetry method regardless of the distance.
  • the image encoding method applied to the present invention may be produced as a program for execution on a computer and stored in a computer-readable recording medium.
  • the computer readable recording medium includes all kinds of storage devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage, and the like, and may also be implemented in the form of a carrier wave (for example, transmission over the Internet). Include.
  • the bitstream generated by the encoding method to which the present invention is applied may be stored in a computer-readable recording medium or transmitted using a wired / wireless communication network.
  • the present invention can be applied to an encoder device and a decoder device for efficient video encoding.
  • the present invention has been described by way of limited embodiments and drawings, the present invention is not limited thereto and is intended by those skilled in the art to which the present invention pertains, and claims to be described below. Various modifications and variations may be made within the scope of equivalents of the scope.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Abstract

La présente invention concerne un procédé et un appareil de traitement de signal vidéo. Le procédé de traitement de signal vidéo selon un mode de réalisation de la présente invention comprend les étapes suivantes consistant à : acquérir des informations de vecteur de mouvement mises à l'échelle pour un bloc à décoder ; acquérir un vecteur de mouvement global d'une image périphérique ; rétablir le vecteur de mouvement mis à l'échelle en un vecteur de mouvement d'origine au moyen du vecteur de mouvement global de l'image périphérique; et décoder le bloc à décoder, au moyen du vecteur de mouvement d'origine rétabli. Selon divers modes de réalisation de la présente invention, un encodage et un décodage d'image peuvent être effectués d'une manière efficace. Par exemple, les distorsions provoquées par un phénomène de flou peuvent être évitées, et les distorsions provoquées par un mouvement d'accélération apparu dans la totalité ou une partie d'une image peuvent être évitées. De plus, des divisions en blocs appropriées pour des caractéristiques d'une image peuvent être réalisées. En particulier, des divisions en blocs au moyen d'une courbe peuvent être réalisées, permettant de ce fait l'exécution d'une prédiction de mouvement en prenant en considération toutes les caractéristiques d'une image. De plus, une zone de recherche d'une taille variable est déterminée proportionnellement à un temps et une distance, évitant de ce fait des distorsions dans un mode de prédiction directe.
PCT/KR2011/000729 2010-02-01 2011-02-01 Procédé et appareil de traitement de signal vidéo Ceased WO2011093688A2 (fr)

Applications Claiming Priority (14)

Application Number Priority Date Filing Date Title
US30003810P 2010-02-01 2010-02-01
US61/300,038 2010-02-01
US31417010P 2010-03-16 2010-03-16
US61/314,170 2010-03-16
US32956210P 2010-04-30 2010-04-30
US61/329,562 2010-04-30
US34515510P 2010-05-17 2010-05-17
US61/345,155 2010-05-17
US34825810P 2010-05-26 2010-05-26
US61/348,258 2010-05-26
US34989210P 2010-05-31 2010-05-31
US61/349,892 2010-05-31
US36142510P 2010-07-04 2010-07-04
US61/361,425 2010-07-04

Publications (2)

Publication Number Publication Date
WO2011093688A2 true WO2011093688A2 (fr) 2011-08-04
WO2011093688A3 WO2011093688A3 (fr) 2012-01-05

Family

ID=44320015

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2011/000729 Ceased WO2011093688A2 (fr) 2010-02-01 2011-02-01 Procédé et appareil de traitement de signal vidéo

Country Status (1)

Country Link
WO (1) WO2011093688A2 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018097589A1 (fr) * 2016-11-22 2018-05-31 한국전자통신연구원 Procédé et dispositif de codage/décodage d'image, et support d'enregistrement sur lequel est stocké un flux binaire
WO2018097590A1 (fr) * 2016-11-22 2018-05-31 한국전자통신연구원 Procédé et dispositif de codage/décodage d'image, et support d'enregistrement sur lequel est stocké un flux binaire

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100446083B1 (ko) * 2002-01-02 2004-08-30 삼성전자주식회사 움직임 추정 및 모드 결정 장치 및 방법
KR20040099100A (ko) * 2002-04-19 2004-11-26 마쯔시다덴기산교 가부시키가이샤 움직임 벡터 계산 방법
JP4130783B2 (ja) * 2002-04-23 2008-08-06 松下電器産業株式会社 動きベクトル符号化方法および動きベクトル復号化方法
KR100800772B1 (ko) * 2004-05-26 2008-02-01 마츠시타 덴끼 산교 가부시키가이샤 움직임 벡터 부호화 장치, 방법, 프로그램 및 매체

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018097589A1 (fr) * 2016-11-22 2018-05-31 한국전자통신연구원 Procédé et dispositif de codage/décodage d'image, et support d'enregistrement sur lequel est stocké un flux binaire
WO2018097590A1 (fr) * 2016-11-22 2018-05-31 한국전자통신연구원 Procédé et dispositif de codage/décodage d'image, et support d'enregistrement sur lequel est stocké un flux binaire

Also Published As

Publication number Publication date
WO2011093688A3 (fr) 2012-01-05

Similar Documents

Publication Publication Date Title
WO2016204360A1 (fr) Procédé et dispositif de prédiction de bloc basée sur la compensation d'éclairage dans un système de codage d'image
JP2022505578A (ja) サブブロックベースのモーション補償を用いたビデオ信号処理方法及び装置
WO2016200043A1 (fr) Procédé et appareil d'inter-prédiction en fonction d'une image de référence virtuelle dans un système de codage vidéo
WO2013069932A1 (fr) Procédé et appareil de codage d'image, et procédé et appareil de décodage d'image
WO2017052000A1 (fr) Procédé et appareil de prédiction inter basée sur le raffinement des vecteurs de mouvement dans un système de codage d'images
WO2013062193A1 (fr) Procédé et appareil de décodage d'image
WO2011149291A2 (fr) Procédé et appareil de traitement d'un signal vidéo
WO2017069419A1 (fr) Procédé et appareil de prédiction intra dans un système de codage vidéo
WO2015142054A1 (fr) Procédé et appareil pour traiter des signaux vidéo multi-vues
WO2018212430A1 (fr) Procédé de filtrage de domaine de fréquence dans un système de codage d'image et dispositif associé
WO2012018197A2 (fr) Dispositif de décodage intra-prédictif
WO2012018198A2 (fr) Dispositif de génération de blocs de prédiction
WO2011145819A2 (fr) Dispositif et procédé de codage/décodage d'image
WO2010087620A2 (fr) Procédé et appareil de codage et de décodage d'images par utilisation adaptative d'un filtre d'interpolation
WO2018199468A1 (fr) Procédé et appareil d'estimation de flux optique à des fins de compensation de mouvement
WO2019112071A1 (fr) Procédé et appareil de décodage d'image basés sur une transformation efficace de composante de chrominance dans un système de codage d'image
WO2013062198A1 (fr) Appareil de décodage d'images
WO2012011672A2 (fr) Procédé et dispositif de codage/décodage d'images par mode de saut étendu
WO2012053796A2 (fr) Appareil et procédé destinés à coder/décoder un vecteur de mouvement différentiel, et appareil et procédé destinés à coder/décoder une vidéo à l'aide de ces derniers
WO2017052272A1 (fr) Procédé et appareil pour une prédiction intra dans un système de codage vidéo
WO2013062194A1 (fr) Procédé et appareil de génération de bloc reconstruit
WO2015152504A1 (fr) Procédé et dispositif pour dériver un candidat à la fusion de mouvement inter-vues
WO2014171709A1 (fr) Procédé et appareil de compensation de luminosité adaptative basés sur objet
WO2011093688A2 (fr) Procédé et appareil de traitement de signal vidéo
WO2010044566A2 (fr) Appareil de codage/décodage de film et procédé et appareil pour compensation du mouvement de blocs se chevauchant et compensation du mouvement de blocs hybrides correspondants

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: 11737327

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 11737327

Country of ref document: EP

Kind code of ref document: A2