WO2014121591A1 - 视频编码质量的评估方法及设备 - Google Patents

视频编码质量的评估方法及设备 Download PDF

Info

Publication number
WO2014121591A1
WO2014121591A1 PCT/CN2013/080882 CN2013080882W WO2014121591A1 WO 2014121591 A1 WO2014121591 A1 WO 2014121591A1 CN 2013080882 W CN2013080882 W CN 2013080882W WO 2014121591 A1 WO2014121591 A1 WO 2014121591A1
Authority
WO
WIPO (PCT)
Prior art keywords
video
content
complexity
frame
slice
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/CN2013/080882
Other languages
English (en)
French (fr)
Inventor
高山
孙李娜
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.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
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 Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to EP13874341.4A priority Critical patent/EP2919461B1/en
Priority to ES13874341.4T priority patent/ES2622027T3/es
Publication of WO2014121591A1 publication Critical patent/WO2014121591A1/zh
Priority to US14/751,373 priority patent/US9774855B2/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102Methods 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
    • H04N19/124Quantisation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N17/00Diagnosis, testing or measuring for television systems or their details
    • H04N17/004Diagnosis, testing or measuring for television systems or their details for digital television systems
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/85Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using pre-processing or post-processing specially adapted for video compression
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/136Incoming video signal characteristics or properties
    • H04N19/14Coding unit complexity, e.g. amount of activity or edge presence estimation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/154Measured or subjectively estimated visual quality after decoding, e.g. measurement of distortion
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169Methods 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
    • H04N19/17Methods 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
    • H04N19/174Methods 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
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/189Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the adaptation method, adaptation tool or adaptation type used for the adaptive coding
    • H04N19/196Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the adaptation method, adaptation tool or adaptation type used for the adaptive coding being specially adapted for the computation of encoding parameters, e.g. by averaging previously computed encoding parameters
    • H04N19/197Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the adaptation method, adaptation tool or adaptation type used for the adaptive coding being specially adapted for the computation of encoding parameters, e.g. by averaging previously computed encoding parameters including determination of the initial value of an encoding parameter

Definitions

  • Video coding quality evaluation method and device
  • Embodiments of the present invention relate to the field of communication technologies, and more particularly, to a method and apparatus for evaluating video coding quality. Background technique
  • Video quality is affected by many factors, such as compression distortion, adaptation of video codec parameters to transport channels, and quality of transport channel services (such as bandwidth limitations, packet loss, delay, or jitter). .
  • the model for predicting the coding quality (the coding quality is also called the reference quality or the compression quality) only considers the coding information (such as the code rate and coding format Codec, video pause and network packet loss).
  • the model of predictive coding quality in the no reference objective video quality assessment method is not comprehensive and does not accurately reflect the subjective feelings of the human eye. It has certain limitations. Therefore, the coding quality prediction Low accuracy. Summary
  • Embodiments of the present invention provide a method and a device for evaluating video coding quality, which can improve the accuracy of coding quality prediction.
  • a method for evaluating video coding quality comprising: acquiring a quantization parameter of a strip of a video frame of a video stream and a pixel number of a strip of a video frame of the video stream; The quantization parameter of the strip of the video frame of the video stream determines the quantization parameter of the video, and the root Determining the content complexity of the video according to the number of pixel bytes of the strip of the video frame of the video stream; predicting the video encoding quality according to the content complexity of the video and the quantization parameter of the video.
  • the determining, by the quantization parameter of the strip of the video frame of the video stream, the quantization parameter of the video may be: using the video frame of the video stream The average or weighted average of the quantization parameters of the strips is determined as the quantization parameter of the video.
  • the determining, by using the average or weighted average of the quantization parameters of the strips of the video frames of the video stream, The quantization parameter of the video may be: determining that the quantization parameter of the video is:
  • f-video_qp is a quantization parameter of the video
  • N1 is a number of quantization parameters of a stripe of the video stream, and is a quantization parameter of an i-th stripe of the quantization parameter of the N1 stripe
  • N1 is a positive integer
  • w is the weight corresponding to the i-th stripe.
  • the pixel according to the strip of the video frame of the video stream The number of bytes determines the content complexity of the video.
  • the specific implementation may be: determining the number of pixel bytes of the kth stripe in the Num strips of the video frame:
  • i- slice- size k is the number of bytes Num bands with the video frame in the k-th bands
  • i- slice_pixel k Num a strip bands of the video frame in the k-th band section
  • the number of pixels, k is a positive integer, in the video frame, the value of k is from 1 to Num, and Num is a positive integer;
  • a [i- slice- qp k] and b [i- slice- qp k] are the video frames of said strip Num band k quantization parameter band corresponding to the value
  • Determining the content complexity of the video frame is:
  • the content complexity of the video is determined according to a content complexity of the video frame, and the specific implementation may be:
  • f-video_content_utility is the content complexity of the video
  • the video stream includes M video frames, M is a positive integer, j is a positive integer and takes values from 1 to M.
  • the video coding quality is predicted according to the content complexity of the video and the quantization parameter of the video, and the specific implementation may be:
  • d-compression-quality_value is the video coding quality
  • aa 2 , a 3 , a 4 , a 5 and a 6 are constants.
  • the specific implementation may be: the value of the 3 ⁇ 4 is: a difference between a maximum value and a minimum value of a video coding quality range.
  • Binding fifth possible implementation of the first aspect or any one sixth possible embodiment of the implementation in a seventh possible implementation, the specific implementation may be: the value of ai: The minimum value of the video encoding quality range.
  • the method further includes: performing content complexity on the video Adjustment: Video content complexity
  • the adjusted f-video_content-utility-n is used as the content complexity of the video, and s and numl are constant.
  • the specific implementation may be: the s takes a value of 1.0, the numl takes a value of 60.0, and the f-vidio- Content—utility—n ranges from [0.0, 1.0].
  • the specific implementation may be: No packet loss.
  • the specific implementation may be: the video frame includes a frame. Intra-coded frames and/or inter-coded frames.
  • the specific implementation may be: Num is the video. The total number of stripes in the frame.
  • the specific implementation may be: a number of quantized parameters of the parsed stripe, the quantized parameter of the correctly parsed stripe representing a stripe quantization parameter difference between the first transport packet of the stripe and the quantization parameter used to obtain the stripe When the transmission packet is not lost, the quantization parameter of the parsed strip is correct.
  • an apparatus for evaluating a video coding quality includes: an acquiring unit, a quantization parameter for acquiring a stripe of a video frame of a video stream, and a pixel of a strip of a video frame of the video stream a determining unit, configured to determine a quantization parameter of the video according to a quantization parameter of a strip of the video frame of the video stream acquired by the acquiring unit, and according to the video frame of the video stream acquired by the acquiring unit
  • the number of pixel bytes of the strip determines the content complexity of the video; the prediction unit is configured to predict the video encoding quality according to the content complexity of the video and the quantization parameter of the video.
  • the determining unit is specifically configured to: An average or weighted average of the quantization parameters of the strips of the video frames of the video stream is determined as the quantization parameter of the video.
  • the determining unit is specifically configured to: determine that the quantization parameter of the video is:
  • N1 is the number of quantization parameters of the stripe of the video stream, and is the quantization parameter of the N1 stripe
  • the quantization parameter of the i-th stripe, N1 is a positive integer, w ; is the weight corresponding to the i-th stripe.
  • the determining unit is specifically configured to:
  • the number of pixel bytes of the kth stripe in the Num stripe of the video frame is determined as:
  • i- slice- size k is the number of bytes Num bands with the video frame in the k-th bands
  • i- slice_pixel k Num a strip bands of the video frame in the k-th band section
  • the number of pixels, k is a positive integer, in the video frame, the value of k is from 1 to Num, and Num is a positive integer;
  • the content complexity of the kth stripe in the Num stripe of the video frame is determined as follows:
  • a [i- slice- qp k] and b [i- slice- qp k] are the quantization parameter corresponding article Num a strip of the video frame in the k-th band values;
  • the content complexity of the video is determined according to content complexity of the video frame.
  • the determining unit is specifically configured to: determine that the content complexity of the video is:
  • the video stream includes M video frames, where M is a positive integer, j is a positive integer, and the value ranges from 1 to M.
  • the a 2 value is: a difference between a maximum value and a minimum value of a video coding quality range.
  • the value of 1 is: a video coding quality range. Minimum value.
  • the method further includes: performing content complexity on the video Adjustment:
  • f _ video _ content _ complexity _ n The adjusted f-video_content-utility-n is used as the content complexity of the video, and both s and numl are constant.
  • the s takes a value of 1.0
  • the numl takes a value of 60.0
  • the f-vidio_content-utility-n The value range is [0.0, 1.0].
  • the M video frames are not lost.
  • the video frame includes an intra-coded frame and/or Or inter-coded frames.
  • Num is a stripe in the video frame. total.
  • the N1 is a correctly parsed strip.
  • the number of quantization parameters, the quantized parameter of the correctly parsed stripe indicates that the transport packet of the stripe quantization parameter difference between the first transport packet of the stripe and the quantization parameter used to obtain the stripe is In the case of no loss, the quantization parameter of the parsed strip is correct.
  • the embodiment of the present invention determines the quantization parameter of the video according to the quantization parameter of the strip of the video frame of the acquired video stream, and determines the content complexity of the video according to the number of pixel bytes of the strip of the video frame of the obtained video stream.
  • the content complexity of the video and the quantization parameters of the video predict the quality of the video encoding. Therefore, the quality of the prediction predicted by the model obtained by considering the characteristics of the video content is more in line with the subjective feeling of the human eye, thereby improving the accuracy of the prediction.
  • FIG. 1 is a flow chart of a method for evaluating video encoding quality according to an embodiment of the present invention.
  • FIG. 2 is a block diagram showing the structure of an apparatus for evaluating video coding quality according to an embodiment of the present invention.
  • FIG. 3 is a block diagram showing the structure of an apparatus for evaluating video coding quality according to another embodiment of the present invention. detailed description
  • GSM Global System for Mobile communications
  • CDMA Code Division Multiple Access
  • WCDMA Wideband Code Division Multiple Access
  • GPRS General Packet Radio Service
  • LTE Long Term Evolution
  • LTE FDD Frequency Division Duplex
  • LTE TDD Time Division Duplex
  • WiMAX Worldwide Interoperability for Microwave Access, global interconnected microwave access
  • FIG. 1 is a flow chart of a method for evaluating video encoding quality according to an embodiment of the present invention. The method is performed by an evaluation device for video coding quality.
  • the quantization parameter may also use another expression method, a quantization step size (QP Step), to perform prediction video coding quality.
  • QP Step a quantization step size
  • the model for predicting video coding quality in the embodiment of the present invention will be exemplified by quantization parameters. It should be understood that it is also within the scope of the present invention to replace the QP expression with QP Step.
  • the embodiment of the present invention determines the quantization parameter of the video according to the quantization parameter of the strip of the video frame of the acquired video stream, and determines the content complexity of the video according to the number of pixel bytes of the strip of the video frame of the obtained video stream.
  • the content complexity of the video and the quantization parameters of the video predict the quality of the video encoding. Therefore, the quality of the prediction predicted by the model obtained by considering the characteristics of the video content is more in line with the subjective feeling of the human eye, thereby improving the accuracy of the prediction.
  • the quantization parameter of the video may be determined according to the quantization parameter of the strip of the video frame of the video stream, and the video may be determined according to the quantization parameter of the macroblock of the video frame of the video stream.
  • the quantization parameter can also be combined with the quantization parameter of the strip of the video frame of the video stream.
  • the quantization parameters of the macroblocks of the video frames of the video and video streams are used to determine the quantization parameters of the video.
  • the total number of quantization parameters of the stripe of the video stream is N
  • the quantization parameters of the N1 stripe may be selected among the quantization parameters of the N strips of the video.
  • the quantization parameter of the N1 strips may be the total number of strips that correctly parse the quantization parameter.
  • the stripe of the quantization parameter is correctly parsed and represented in the image parameter set (pic_parameter_set_rbsp) , PPS) image initial quantization minus 26 (pic_init_qp_minus26, pic_init_qp_minus26 is a syntax element in the H.264 image parameter set) where the transport packet is not lost, the first transport packet of the stripe is used to obtain the The quantization parameter difference of the stripe quantization parameter strip (slice_qp-delta, slice_qp-delta is a syntax element in the H.264 strip header) is not lost; in other words, The first bit of the band is not lost to the bit of the quantization parameter used to resolve the stripe (slice_qp-delta, slice_qp-delta is a syntax element in the H.264 strip header) The quantization parameter difference of the band is used to obtain the quantization parameter of the strip; thus, the quantization parameter of the strip can be correctly parsed.
  • the quantization parameter difference of the band is used
  • N is a positive integer
  • Nl is a positive integer
  • the average value of the quantization parameters of the N1 strips may be determined as a quantization parameter of the above video for predicting the encoding quality, as the quantization parameter of the following (1) video is f-video_qp:
  • i_slice_9 1 is a quantization parameter of the i-th stripe in the quantization parameter of 1 ⁇ 1 stripe.
  • the foregoing N1 may also be the number of quantization parameters of the stripe of the stripe that does not occur.
  • the weighted average of the quantization parameters of the N1 strips may be determined as a quantization parameter of the video used for predicting the encoding quality, w ; the weight corresponding to the ith strip may be as needed Different weights are given to different stripes. As shown in the following formula:
  • the quantization parameter f_video_qp of the video obtained by the above equations may be normalized and adjusted, and the adjusted quantization parameter of the video is used for predicting the coding quality. It should be understood that the above examples are merely exemplary. The embodiments of the present invention do not limit how to obtain the quantization parameters of the video used for predicting the encoding quality, and may also correctly parse all intra-coded frames and inter-coded frames in the video.
  • the average or weighted average of the quantization parameters of the stripe is determined as the quantization parameter of the above video for predicting the encoding quality, and the average or weighted average of the quantization parameters of all the strips and macroblocks in the video may also be determined as being used for A quantization parameter that predicts the quality of the video. and many more.
  • a packet loss may be generated, and a packet may be selected for the packet loss or the packet without the packet loss, and the embodiment of the present invention is not limited thereto.
  • the video frame may be an inter-coded frame and/or an intra-coded frame
  • the inter-coded frame may be an inter-coded frame (also referred to as an inter-frame reference frame), such as a P frame, as a reference frame.
  • B frame may also be an inter-coded frame that is not a reference frame, such as a b-frame
  • an intra-coded frame may be an I-frame or an IDR (Instantaneous Decoding Refresh) frame, also referred to as an intra-frame reference frame. and many more. It should be understood that the embodiments of the present invention are not limited thereto.
  • the content complexity of the video is determined according to the number of pixel bytes of the strip of the video frame of the video stream.
  • the content complexity of the video frame is determined according to the number of pixel bytes of the strip of the video frame of the video stream, and then according to the video.
  • the content complexity of the frame determines the content complexity of the video, specifically, as shown in the following (2) - (4):
  • the number of pixel bytes of the kth stripe in the Num stripe of the video frame is:
  • i- slice- size k is the number of bytes Num bands with the video frame in the k-th bands
  • i- slice_pixel k Num a strip bands of the video frame in the k-th band section
  • the number of pixels, k is a positive integer.
  • k takes a value from 1 to Num
  • Num is a positive integer.
  • the number of pixel bytes may be modified.
  • the number of pixel bits may be deformed, and then the corresponding operation is performed by using the number of pixel bits.
  • the deformation does not deviate from the embodiment of the present invention. Substantially, it should still fall within the scope of protection of the present invention.
  • the content complexity of the kth stripe in the Num strips of the video frame is:
  • a [i- slice- qp k] and b [i- slice- qp k] corresponding to quantization parameters are with a strip Num of the video frame in the value of k stripes.
  • the content complexity of a video frame is:
  • Num may be the number of all the strips of the video frame (including the number of lost packets and the number of strips that are not lost), and Num may also be the number of packets that are not lost in the video frame or packet loss.
  • the number of strips is not limited in this embodiment of the invention.
  • the average or weighted average of the content complexity of the multiple video frames may be determined as the content of the video for predicting the encoding quality.
  • Degree specifically, as shown in the following formula (5):
  • the content complexity of a video frame is:
  • the video stream includes M video frames
  • M is a positive integer
  • j is a positive integer and takes values from 1 to M.
  • none of the M video frames are lost.
  • the complexity of the content of the video may include the time direction and/or the spatial direction. Under a certain fixed quantization parameter of a certain resolution, the greater the complexity of the content of the video, the less the distortion is easily found by the human eye.
  • the complexity of the video content obtained by the above technical solution is more complicated than the complexity of the video content, and the distortion is less likely to be subjectively perceived by the human eye, thereby further improving the accuracy of the prediction coding quality.
  • the video frame in the above (2) - (5) is an intra reference frame, that is, the content complexity of the video can be determined by the number of stripe pixel bytes of the intra reference frame (using at least one intra reference frame)
  • the content complexity of the video determined by the number of stripe pixel bytes of the intra reference frame can effectively reflect the spatial complexity of the video f-video_content-utility-sec
  • the sub-a [i- slice- qp k] and b [i- slice- qp k] can be obtained through training, different video resolution and / or encoding formats and / or video capture format (i.e.: an interlace mode and Progressive scan mode) You can use the same value or a different value.
  • the video frames in the above (2) - (5) are inter-frame reference frames, that is, the content complexity of the video may be determined by the number of stripe pixel bytes of the inter-frame reference frame (using at least one inter-frame reference frame)
  • the content complexity of the video determined by the number of stripe pixel bytes of the interframe reference frame can effectively reflect the time complexity of the video f_video_content-utility-tec, optionally, the expression a [i- slice- qp k] and b [i- slice- qp k] can be obtained through training, different views
  • the frequency resolution and/or the encoding format and/or the video capture format may use the same value or different values.
  • the quantization parameter 52 having (0-51, respectively), for example, respectively, obtained through a training at different resolutions [i- slice- qp k] and b [i- slice- qp k ] , corresponding to the quantization parameter values 0 to 51, respectively.
  • the content complexity of the video according to the above (5) may be f_video_content-utility and the amount of the video of the above (1)
  • the predicted video encoding quality is:
  • the values up to & 6 may be the same or different under different video resolutions and/or encoding formats and/or video capture formats (ie, interlaced and progressive).
  • the value of 1 to a 6 in the standard definition and the value of the full HD resolution ⁇ to a 6 may be the same or different. It should be understood that the embodiment of the present invention does not limit this.
  • the video quality is better, and the smaller the value of the video coding quality d-compression-quality-value, the worse the video quality is.
  • the value of a 2 may be the difference between the maximum value and the minimum value of the video coding quality range.
  • the code quality has a maximum value of 5 and a minimum value of 1, so a 2 takes a value of 4.
  • the value of a 2 may be the difference between the maximum value and the minimum value of the video coding quality range obtained through training or experience, such as obtaining a maximum coding quality of 4.5 by training or experience, and a minimum coding quality of 1.5.
  • the value of a 2 is 3. It should be understood that the manner in which the encoding quality is predicted in the embodiment of the present invention may be applied to the metric of the encoding quality of other metrics, which is used by the embodiment of the present invention. No restrictions.
  • the value of a may be a minimum value of the video coding quality range.
  • the values of a 5 and a 6 may also be determined according to different situations. It should be understood that the embodiment of the present invention does not limit this.
  • the values of ai , a 2 , a 3 , a 4 , a 5 and a 6 can also be trained or empirically valued according to different situations.
  • the manner of expressing the above-mentioned coding quality by other equivalent formulas falls within the scope of the present invention.
  • the video coding distortion quality d-compression-artificial value is subtracted from the maximum value of the video coding quality by the above video coding.
  • the quality obtained is as shown in the following formula (7):
  • the content complexity of the video in the above formulas (5) to (7) may be adjusted, and the adjusted f-video_content_utility-n is used as the content complexity of the video, and the adjustment manner is adopted.
  • the adjustment manner is adopted.
  • s and numl are constants. For example, when performing normalization adjustment, s takes a value of 1.0, numl takes a value of 60.0, and f_vidio-content-utility-n takes a value of [0.0, 1.0].
  • the coding quality or the video coding distortion quality is predicted by the formula (6) or (7)
  • the content complexity of the video adjusted by the equation (8) may be adopted, that is, the prediction video coding.
  • the quality can be:
  • the predicted video encoding quality can also be:
  • the content complexity of the adjusted video f-video-content-utility-n may be the time complexity of the normalized adjusted video, f-video-content-utility-ntcc, or normalized.
  • the spatial complexity of the adjusted video f-video-content-utility-nscc can also be a combination of time complexity f-video-content-utility-ntcc and space complexity f-video-content-utility-nscc. It should be understood that the embodiments of the present invention do not limit this.
  • time complexity f-video-content-utility-ntcc and space complexity f-video-content-utility- nscc combination and video quantity 4 ⁇ parameter f video qp prediction video coding quality d-compression-quality- Value that is, the complexity of the content of the video in (6) adopts the complexity of the content of the adjusted video f-video-content-utility-n, a 5 *f— video—content— complexity—n replaced by a 51 *f— video—content— complexity— ntcc+a 52 *f— video—content—physical— nscc , where a 51 and a 52 may be the same or different.
  • the method for predicting the coding quality by the embodiment of the present invention is more in line with the subjective feeling of the human eye, thereby further improving the accuracy of the prediction coding quality.
  • the content complexity of the video can be determined according to the coding information of other video frames, such as the number of encoded bytes of the video frame, the motion vector difference, the motion vector value, the macroblock coding mode, and the DCT (Discrete Cosine Transform, Discrete cosine transform) coefficients, etc.
  • the coding information of other video frames such as the number of encoded bytes of the video frame, the motion vector difference, the motion vector value, the macroblock coding mode, and the DCT (Discrete Cosine Transform, Discrete cosine transform) coefficients, etc.
  • the content complexity of a video is expressed as the average of the AC (alternating current) coefficients of the discrete cosine transform DCT coefficients of all intra-coded frames (I/IDR frames), or all intra-coded frames (I/IDR frames)
  • the average of the DC (Direct Current) coefficients of the discrete cosine transform DCT coefficients, or the average of the discrete cosine transform DCT coefficients of all intra-coded frames (I/IDR frames), or the dispersion of inter-coded frames The average of the AC coefficients of the cosine transform DCT coefficients, or the average of the DC coefficients of the discrete cosine transform DCT coefficients of the interframe coded frame, or the average of the discrete cosine transform DCT coefficients of the interframe coded frame, or the intraframe coded frames and frames.
  • the content complexity of the video is described by a macroblock coding mode (e.g., 16x16, 16x8, 8x16, etc.), and the more the macroblock coding mode, the higher the content complexity of the video.
  • the content complexity of a video is described by a motion vector. The larger the motion vector, the higher the complexity of the video content.
  • the content complexity of the video is represented by the ratio of the inter-coded block to the intra-coded block. The larger the ratio, the higher the content complexity of the video. and many more. It should be understood that the embodiments of the present invention are not limited thereto.
  • the video coding quality evaluation apparatus 200 includes an acquisition unit 201, a determination unit 202, and a prediction unit 203.
  • the obtaining unit 201 is configured to obtain a quantization parameter of a stripe of a video frame of the video stream, and a number of pixel bytes of a stripe of the video frame of the video stream.
  • the determining unit 202 is configured to determine a quantization parameter of the video according to the quantization parameter of the strip of the video frame of the video stream acquired by the acquiring unit 201, and determine, according to the number of pixel bytes of the strip of the video frame of the video stream acquired by the obtaining unit 201.
  • the content complexity of the video is configured to determine a quantization parameter of the video according to the quantization parameter of the strip of the video frame of the video stream acquired by the acquiring unit 201, and determine, according to the number of pixel bytes of the strip of the video frame of the video stream acquired by the obtaining unit 201. The content complexity of the video.
  • the prediction unit 203 is configured to predict a video coding quality according to the content complexity of the video determined by the determining unit 202 and the quantization parameter of the video.
  • the quantization parameter may also adopt another expression manner, and the quantization step size (QP) Step) to predict the quality of video encoding.
  • QP quantization step size
  • the model for predicting video coding quality in the embodiment of the present invention will be exemplified by quantization parameters. It should be understood that it is also within the scope of the present invention to replace the QP expression with QP Step.
  • the embodiment of the present invention determines the quantization parameter of the video according to the quantization parameter of the strip of the video frame of the acquired video stream, and determines the content complexity of the video according to the number of pixel bytes of the strip of the video frame of the obtained video stream.
  • the content complexity of the video and the quantization parameters of the video predict the quality of the video encoding. Therefore, the quality of the prediction predicted by the model obtained by considering the characteristics of the video content is more in line with the subjective feeling of the human eye, thereby improving the accuracy of the prediction.
  • the device 200 can implement the steps of the evaluation device involving the video coding quality in the method of Fig. 1. To avoid repetition, it will not be described in detail.
  • the determining unit 202 may be specifically configured to: determine an average value or a weighted average value of the quantization parameters of the strips of the video frames of the video stream as the quantization parameter of the video.
  • the determining unit 202 can be used to:
  • N1 is the number of quantization parameters of the stripe of the video stream, i-slicc_qp ⁇ the quantization parameter of the i-th stripe in the quantization parameter of the N1 stripe, N1 is a positive integer, w ; is the ith The weight corresponding to the strip.
  • the N1 is the total number of quantization parameters of the correctly parsed strip
  • the correctly parsed quantization parameter of the strip indicates the strip quantization of the first transport packet in the strip to the quantization parameter used to obtain the stripe.
  • the quantization parameter of the parsed strip is correct.
  • the determining unit 202 may further perform normalization adjustment on the quantization parameter f_video_qp of the video obtained in the above formula, and use the quantization parameter of the adjusted video for predicting the encoding quality.
  • the determining unit 202 may be specifically configured to:
  • the number of pixel bytes of the kth stripe in the Num stripe of the video frame is determined as: i slice size.
  • the number of bytes Num i- slice- size k is a strip of a video frame of the k-th bands, the number of pixels Num i- slice_pixel k th slice of a video frame of the k-th bands, k
  • k takes values from 1 to Num, and Num is a positive integer
  • a [i- slice- qp k] corresponding to quantization parameter values b [i- slice- qp k] are Num a strip article with a video frame in the k-th zone;
  • it is used to: determine the content complexity of the video according to the content complexity of the video frame.
  • the video frame may be an inter-coded frame and/or an intra-coded frame
  • the inter-coded frame may be an inter-coded frame (also referred to as an inter-frame reference frame), such as a P frame, as a reference frame.
  • B frame may also be an inter-coded frame that is not a reference frame, such as a b-frame
  • an intra-coded frame may be an I-frame or an IDR (Instantaneous Decoding Refresh) frame, also referred to as an intra-frame reference frame. and many more. It should be understood that the embodiments of the present invention are not limited thereto.
  • the determining unit may be specifically configured to: determine that the content complexity of the video is:
  • the video stream includes M video frames, M is a positive integer, j is a positive integer and takes values from 1 to M.
  • Num can be the number of all the strips of the video frame (including the number of lost and undelivered strips), and Num can also be the number of non-dropped strips of the video frame or the stripe of the lost packet.
  • the embodiments of the present invention are not limited thereto.
  • the video frame may comprise an intra-coded frame and/or an inter-coded frame.
  • the content complexity of the video may include the time direction and/or the spatial direction. Under a certain fixed quantization parameter of a certain resolution, the greater the complexity of the content of the video, the less the distortion is easily found by the human eye.
  • the complexity of the video content obtained by the above technical solution is more in line with the complexity of the video content, and the distortion is less likely to be subjectively perceived by the human eye, thereby further improving the accuracy of the prediction encoding quality.
  • the prediction unit 203 may be specifically configured to: predict the video coding quality as:
  • the video quality is better, and the smaller the value of the video coding quality d-compression-quality-value, the worse the video quality is.
  • the value of a 2 may be the difference between the maximum value and the minimum value of the video coding quality range.
  • the code quality has a maximum value of 5 and a minimum value of 1, so a 2 takes a value of 4.
  • the value of a 2 may be the difference between the maximum value and the minimum value of the video coding quality range obtained through training or experience, such as obtaining a maximum coding quality of 4.5 by training or experience, and a minimum coding quality of 1.5.
  • the value of a 2 is 3. It should be understood that the method for predicting the quality of the coding in the embodiment of the present invention may be applied to the metric of the coding quality of other systems, which is not limited by the embodiment of the present invention.
  • the value of a may be a minimum value of the video coding quality range.
  • the values of a 5 and a 6 may also be determined according to different situations. It should be understood that the embodiment of the present invention does not limit this.
  • the values of ai , a 2 , a 3 , a 4 , a 5 and a 6 can also be trained or empirically valued according to different situations.
  • the manner of expressing the above-mentioned coding quality by other equivalent formulas falls within the scope of the present invention.
  • the video coding distortion quality d-compression-artificial value is subtracted from the maximum value of the video coding quality by the above video coding.
  • the quality obtained, as in the above (7), at this time the smaller the video coding distortion quality, the better the video quality, and the greater the video coding distortion quality, the worse the video quality.
  • the determining unit 202 may be specifically configured to adjust the content complexity of the video to: Video content complexity
  • s and numl are constants. For example, when performing normalization adjustment, s takes a value of 1.0, numl takes a value of 60.0, and f_vidio-content-utility-n takes a value of [0.0, 1.0].
  • the content complexity of the video adjusted by the (8) formula can be adopted, and the adjusted video is
  • the content complexity f_video_content-utility-n can be the time complexity of the normalized adjusted video f-video-content-utility-ntcc, or the spatial complexity of the normalized adjusted video.
  • F-video—content—physical—nscc can also be a combination of time complexity f—video—content—physical—ntcc and space complexity f—video—content—physical—nscc. It should be understood that the embodiments of the present invention do not limit this.
  • time complexity f-video-content-utility-ntcc and space complexity f-video-content-utility- nscc combination and video quantity 4 ⁇ parameter f video qp prediction video coding quality d-compression-quality- Value that is, the complexity of the content of the video in (6) adopts the complexity of the content of the adjusted video f-video_content-utility-n, which can replace a 5 *f-video-content-utility-n to make a 51 *f - video - content - complexity - ntcc + a 52 *f - video - content - complexity - nscc , where a 51 and a 52 may be the same or different.
  • the method for predicting the coding quality by the embodiment of the present invention is more in line with the subjective feeling of the human eye, thereby further improving the accuracy of the prediction coding quality.
  • FIG. 3 is a structural block diagram of an apparatus for evaluating a video encoding quality according to another embodiment of the present invention.
  • at least one processor 410 such as a CPU
  • at least one port 420 memory 430
  • at least one communication bus 440 are generally included.
  • Communication bus 440 is used to implement such a computer program; alternatively, device 300 can include a user interface 450 including, but not limited to, a display, a keyboard and a pointing device such as a mouse, trackball, touchpad or Touch screen.
  • the memory 430 may include a high speed RAM memory and may also include a non-volatile memory such as at least one disk memory.
  • memory 430 stores the following elements, executable modules or data structures, or a subset thereof, or their extended set.
  • Operating system 432 which contains various system programs for implementing various basic services and handling hardware-based tasks.
  • Application module 434 which contains various applications for implementing various application services.
  • the application module 434 includes, but is not limited to, an obtaining unit 201, a determining unit 202, and a prediction unit.
  • each unit in the application module 434 refers to the corresponding unit in the embodiment shown in FIG. 2, and details are not described herein.
  • the above-mentioned video encoding quality evaluation device may be a terminal, for example, a portable, portable, handheld, computer built-in or in-vehicle mobile device, or the device may be a server. and many more.
  • the disclosed systems, devices, and methods may be implemented in other ways.
  • the device embodiments described above are merely illustrative.
  • the division of the unit is only a logical function division.
  • there may be another division manner for example, multiple units or components may be combined or Can be integrated into another system, or some features can be ignored, or not executed.
  • the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be electrical, mechanical or otherwise.
  • the units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the solution of the embodiment.
  • each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the functions, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium.
  • the technical solution of the present invention which is essential to the prior art or part of the technical solution, may be embodied in the form of a software product stored in a storage medium, including
  • the instructions are used to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention.
  • the foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and the like, which can store program codes. .

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • General Health & Medical Sciences (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Abstract

本发明实施例提供一种视频编码质量的评估方法及设备。该方法包括:获取视频流的视频帧的条带的量化参数,和所述视频流的视频帧的条带的像素字节数;根据所述视频流的视频帧的条带的量化参数确定视频的量化参数,并根据所述视频流的视频帧的条带的像素字节数确定视频的内容复杂度;根据所述视频的内容复杂度和所述视频的量化参数预测视频编码质量。本发明预测视频编码质量还考虑了视频内容复杂度,因此,通过考虑了视频的内容复杂度得到的模型所预测出的编码质量更符合人眼主观感受,从而提高预测的准确性。

Description

视频编码质量的评估方法及设备
本申请要求于 2013 年 2 月 6 日提交中国专利局、 申请号为 201310048015.0、 发明名称为"视频编码质量的评估方法及设备"的中国专利 申请的优先权, 其全部内容通过引用结合在本申请中。 技术领域
本发明实施例涉及通信技术领域, 并且更具体地, 涉及视频编码质量的 评估方法及设备。 背景技术
随着网络技术的发展, 影视点播、 网络电视、 可视电话等视频业务已成 为带宽网络的主要业务, 也将成为 3G ( the 3rd Generation, 第三代)无线网 络的主要业务。
由于视频业务数量大、 实时性高, 以及用户对视频业务敏感性强, 使得 运营商需要对传输的视频业务质量进行监控并及时采取相应措施进行调节 以满足用户对视频业务的体验需求。 视频质量受到很多因素的影响, 例如, 压缩失真, 视频编解码端参数与传输信道的适配情况, 以及传输信道服务质 量(如带宽的限制、 丟包、 时延或抖动等)等。。
现有的无参考客观视频质量评估方法中, 预测编码质量(编码质量也称 为基准质量或压缩质量 )的模型只考虑编码信息(如码率和编码格式 Codec、 视频停顿和网络丟包的影响。 该无参考客观视频质量评估方法(no reference objective video quality assessment method ) 中预测编码质量的模型考虑不全 面, 并不能准确反映人眼的主观感受, 具有一定的局限性, 因此, 编码质量 预测的准确性低。 发明内容
本发明实施例提供一种视频编码质量的评估方法及设备, 能够提高编码 质量预测的准确性。
第一方面, 提供了一种视频编码质量的评估方法, 该方法包括: 获取视 频流的视频帧的条带的量化参数和所述视频流的视频帧的条带的像素字节 数; 根据所述视频流的视频帧的条带的量化参数确定视频的量化参数, 并根 据所述视频流的视频帧的条带的像素字节数确定视频的内容复杂度; 根据所 述视频的内容复杂度和所述视频的量化参数预测视频编码质量。
结合第一方面, 在第一种可能的实现方式中, 所述根据所述视频流的视 频帧的条带的量化参数确定视频的量化参数, 具体实现可以为: 将所述视频 流的视频帧的条带的量化参数的平均值或加权平均值确定为所述视频的量 化参数。
结合第一方面的第一种可能的实现方式, 在第二种可能的实现方式中, 所述将所述视频流的视频帧的条带的量化参数的平均值或加权平均值确定 为所述视频的量化参数, 具体实现可以为: 确定所述视频的量化参数为:
N I
^ i _ slice _ qpt
f video qp=— , 或者
N\
f — video— qp:
Figure imgf000004_0001
其中, f— video— qp为所述视频的量化参数, Nl为所述视频流的条带的量 化参数的数目, 为所述 N1个条带的量化参数中第 i个条带的 量化参数, N1为正整数, w,为该第 i个条带所对应的权重。
结合第一方面或第一方面的第一种可能的实现方式或第二种可能的实 现方式, 在第三种可能的实现方式中, 所述根据所述视频流的视频帧的条带 的像素字节数确定视频的内容复杂度,具体实现可以为: 确定视频帧的 Num 个条带中的第 k个条带的像素字节数为:
i slice size.
f _ slice _ byte _ per _ pixelt
i—slice _pixelk
其中, i— slice— sizek为所述视频帧的 Num个条带中第 k个条带的字节数, i— slice_pixelk为所述视频帧的 Num个条带中第 k个条带的像素数, k为正整 数, 在所述视频帧中, k的取值从 1到 Num,Num为正整数;
确定所述视频帧的 Num个条带中的第 k个条带的内容复杂度为: f _ slice _ content _ complexity k =
a i _ slice _gpk] - f _ slice _ byte _ per _ pixel k + b\i_ slice _ qpt 其中, a[i— slice— qpk]和 b[i— slice— qpk]均为与所述视频帧的 Num个条带中 k个条带的量化参数对应的值;
确定所述视频帧的内容复杂度为:
Num
∑ / _ slice _ content _ complexity f _ frame _ content _ complexity =
Num
根据所述视频帧的内容复杂度确定所述所述视频的内容复杂度。
结合第一方面的第三种可能的实现方式, 在第四种可能的实现方式中, 所述根据所述视频帧的内容复杂度确定所述视频的内容复杂度,具体实现可 以为:
M
_ frame _ content _ complexity j f _ video _ content _ complexity =
M
其中, f— video— content— complexity为所述视频的内容复杂度, 所述视频 流包括 M个视频帧, M为正整数, j为正整数且取值从 1到 M。
结合第一方面的第四种可能的实现方式, 在第五种可能的实现方式中, 所述根据所述视频的内容复杂度和所述视频的量化参数预测视频编码质量, 具体实现可以为:
d _ compression _ quality _ value =
a2 ai + a3 + (一 f -vide0-qP
α4 - 5 · f _ video _ content _ complexity
其中, d— compression— quality— value 为所述视频编码质量, a a2, a3, a4 , a5和 a6均为常数。
结合第一方面的第五种可能的实现方式, 在第六种可能的实现方式中, 具体实现可以为:所述 ¾取值为:视频编码质量范围的最大值与最小值之差。
结合第一方面的第五种可能的实现方式或第六种可能的实现方式中的 任一种方式中, 在第七种可能的实现方式中, 具体实现可以为: 所述 ai取值 为: 视频编码质量范围的最小值。
结合第一方面的第三种可能的实现方式至第七种可能的实现方式中的 任一种方式中, 在第八种可能的实现方式中, 还包括, 对所述视频的内容复 杂度进行调整: video content complexity
Figure imgf000006_0001
将调整后的 f— video— content— complexity— n作为所述视频的内容复杂度, s和 numl均为常数。
结合第一方面的第八种可能的实现方式, 在第九种可能的实现方式中, 具体实现可以为: 所述 s 取值为 1.0 , 所述 numl 取值为 60.0 , 所述 f—vidio— content— complexity— n取值范围为 [0.0, 1.0]。
结合第一方面的第四种可能的实现方式或第九种可能的实现方式中的 任一种方式中, 在第十种可能的实现方式中, 具体实现可以为: 所述 M个 视频帧均没有丟包。
结合第一方面的第四种可能的实现方式或第十种可能的实现方式中的 任一种方式中, 在第十一种可能的实现方式中, 具体实现可以为: 所述视频 帧包括帧内编码帧和 /或帧间编码帧。
结合第一方面的第三种可能的实现方式或第十一种可能的实现方式中 的任一种方式中, 在第十二种可能的实现方式中, 具体实现可以为: Num为 所述视频帧中条带的总数。
结合第一方面的第二种可能的实现方式或第十二种可能的实现方式中 的任一种方式中, 在第十三种可能的实现方式中, 具体实现可以为: 所述 N1 为正确解析出的条带的量化参数的数目, 所述正确解析出的条带的量化 参数表示在条带的第一个传输包至用于获取所述条带的量化参数的条带量 化参数差量所在的传输包都不丟失的情况下,解析出的条带的量化参数为正 确的。
第二方面, 提供了一种视频编码质量的评价设备, 所述设备包括: 获取 单元, 用于获取视频流的视频帧的条带的量化参数和所述视频流的视频帧的 条带的像素字节数; 确定单元, 用于根据所述获取单元获取的所述视频流的 视频帧的条带的量化参数确定视频的量化参数, 并根据所述获取单元获取的 所述视频流的视频帧的条带的像素字节数确定视频的内容复杂度; 预测单 元, 用于根据所述视频的内容复杂度和所述视频的量化参数预测视频编码质 量。
结合第二方面, 在第一种可能的实现方式中, 所述确定单元具体用于: 将所述视频流的视频帧的条带的量化参数的平均值或加权平均值确定为所 述视频的量化参数。
结合第二方面的第一种可能的实现方式, 在第二种可能的实现方式中, 所述确定单元具体用于: 确定所述视频的量化参数为:
N I
^ i _ slice _ qpt
f video qp=— , 或者
N\
N l
^ wt · i _ slice _ qpt
f _ video _ qp = ^ i=l 其中, f— video— qp为所述视频的量化参数, Nl为所述视频流的条带的量 化参数的数目, 为所述 N1个条带的量化参数中第 i个条带的 量化参数, N1为正整数, w;为该第 i个条带所对应的权重。
结合第二方面或第二方面的第一种可能的实现方式或第二种可能的实 现方式, 在第三种可能的实现方式中, 所述确定单元具体用于:
确定视频帧的 Num个条带中的第 k个条带的像素字节数为:
i slice size.
f _ slice _ byte _ per _ pixel}
i—slice—pixelk
其中, i— slice— sizek为所述视频帧的 Num个条带中第 k个条带的字节数, i— slice_pixelk为所述视频帧的 Num个条带中第 k个条带的像素数, k为正整 数, 在所述视频帧中, k的取值从 1到 Num,Num为正整数;
具体用于: 确定所述视频帧的 Num个条带中的第 k个条带的内容复杂 度为:
f _ slice _ content _ complexity k
(i i _ slice _qpk ' f— slice _ byte _ per _ pixelk + b i_ slice _ qpk
其中, a[i— slice— qpk]和 b[i— slice— qpk]均为与所述视频帧的 Num个条带中 的第 k个条带的量化参数对应的值;
具体用于: 确定所述视频帧的内容复杂度为: Num
∑ / _ slice _ content _ complexity f _ frame _ content _ complexity =
Num
具体用于: 根据所述视频帧的内容复杂度确定所述视频的内容复杂度。 结合第二方面的第三种可能的实现方式, 在第四种可能的实现方式中, 所述确定单元具体用于: 确定所述视频的内容复杂度为:
M
frame _ content _ complexity j f _ video _ content _ complexity =
M
其中, 所述视频流包括 M个视频帧, M为正整数, j为正整数且取值从 1到 M。
结合第二方面的第四种可能的实现方式, 在第五种可能的实现方式中, 所述预测单元具体用于: 预测视频编码质量为: d _ compression _ quality _ value =
a2 ai + a3 + (一 f -vide0-qP
a4— a5 · f— video _ content _ complexity
其中, a5和 a6均为常数。
结合第二方面的第五种可能的实现方式, 在第六种可能的实现方式中, 所述 a2取值为: 视频编码质量范围的最大值与最小值之差。
结合第二方面的第五种可能的实现方式或第六种可能的实现方式中的 任一种方式中, 在第七种可能的实现方式中, 所述 1取值为: 视频编码质量 范围的最小值。
结合第二方面的第三种可能的实现方式至第七种可能的实现方式中的 任一种方式中, 在第八种可能的实现方式中, 还包括, 对所述视频的内容复 杂度进行调整:
f _ video _ content _ complexity _ n =
Figure imgf000008_0001
将调整后的 f— video— content— complexity— n作为所述视频的内容复杂度, s和 numl均为常数。 结合第二方面的第八种可能的实现方式, 在第九种可能的实现方式中, 所述 s取值为 1.0,所述 numl取值为 60.0,所述 f—vidio— content— complexity— n 取值范围为 [0.0, 1.0]。
结合第二方面的第四种可能的实现方式或第九种可能的实现方式中的 任一种方式中, 在第十种可能的实现方式中, 所述 M个视频帧均没有丟包。
结合第二方面的第四种可能的实现方式或第十种可能的实现方式中的 任一种方式中, 在第十一种可能的实现方式中, 所述视频帧包括帧内编码帧 和 /或帧间编码帧。
结合第二方面的第三种可能的实现方式或第十一种可能的实现方式中 的任一种方式中,在第十二种可能的实现方式中, Num为所述视频帧中条带 的总数。
结合第二方面的第二种可能的实现方式或第十二种可能的实现方式中 的任一种方式中, 在第十三种可能的实现方式中, 所述 N1为正确解析出的 条带的量化参数的数目, 所述正确解析出的条带的量化参数表示在条带的第 一个传输包至用于获取所述条带的量化参数的条带量化参数差量所在的传 输包都不丟失的情况下, 解析出的条带的量化参数为正确的。
本发明实施例通过根据获取的视频流的视频帧的条带的量化参数确定 视频的量化参数, 并根据获取的视频流的视频帧的条带的像素字节数确定视 频的内容复杂度,通过视频的内容复杂度和视频的量化参数预测视频编码质 量。 因此, 通过考虑了视频内容特性得到的模型所预测出的编码质量更符合 人眼主观感受, 从而提高预测的准确性。 附图说明
为了更清楚地说明本发明实施例的技术方案, 下面将对实施例或现有技 术描述中所需要使用的附图作筒单地介绍, 显而易见地, 下面描述中的附图 仅仅是本发明的一些实施例, 对于本领域普通技术人员来讲, 在不付出创造 性劳动的前提下, 还可以根据这些附图获得其他的附图。
图 1是本发明一个实施例的视频编码质量的评估方法的流程图。
图 2是本发明一个实施例的视频编码质量的评估设备的结构框图。
图 3是本发明另一个实施例的视频编码质量的评估设备的结构框图。 具体实施方式
下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进行 清楚、 完整地描述, 显然, 所描述的实施例是本发明一部分实施例, 而不是 全部的实施例。 基于本发明中的实施例, 本领域普通技术人员在没有作出创 造性劳动前提下所获得的所有其他实施例, 都属于本发明保护的范围。
应理解,本发明实施例的技术方案可以应用于各种通信系统,例如: GSM 系统、 CDMA ( Code Division Multiple Access, 码分多址) 系统、 WCDMA ( Wideband Code Division Multiple Access, 宽带码分多址) 系统、 GPRS ( General Packet Radio Service, 通用分组无线业务)、 LTE系统、 LTE FDD ( Frequency Division Duplex , 频分双工) 系统、 LTE TDD ( Time Division Duplex , 时分双工)、 UMTS , WiMAX ( Worldwide Interoperability for Microwave Access, 全球互联微波接入)通信系统等。
图 1是本发明一个实施例的视频编码质量的评估方法的流程图。该方法 由视频编码质量的评估设备执行。
101 , 获取视频流的视频帧的条带的量化参数, 和视频流的视频帧的条 带的像素字节数。
102, 根据视频流的视频帧的条带的量化参数确定视频的量化参数, 并 根据视频流的视频帧的条带的像素字节数确定视频的内容复杂度。
103 , 根据视频的内容复杂度和视频的量化参数预测视频编码质量。 在本发明实施例中, 量化参数( Quantization Parameter, QP )也可以采 用另一种表达方式, 量化步长(QP Step )来进行预测视频编码质量。 为了描 述方便, 本发明实施例预测视频编码质量的模型中将以量化参数来举例说 明, 应理解, 以 QP Step代替 QP的表达方式也落入本发明范围内。
本发明实施例通过根据获取的视频流的视频帧的条带的量化参数确定 视频的量化参数, 并根据获取的视频流的视频帧的条带的像素字节数确定视 频的内容复杂度,通过视频的内容复杂度和视频的量化参数预测视频编码质 量。 因此, 通过考虑了视频内容特性得到的模型所预测出的编码质量更符合 人眼主观感受, 从而提高预测的准确性。
可选地, 作为一个实施例, 在步骤 102中, 可以根据视频流的视频帧的 条带的量化参数确定视频的量化参数,还可以根据视频流的视频帧的宏块的 量化参数确定视频的量化参数,也可以结合视频流的视频帧的条带的量化参 数和视频流的视频帧的宏块的量化参数来确定视频的量化参数。
具体地,视频流的条带的量化参数的总数为 N,可以在视频的 N个条带 的量化参数中选择 N1个条带的量化参数。 优选地, 该 N1个条带的量化参 数可以是正确解析出量化参数的条带的总数, 以 H.264为例, 正确解析出量 化参数的条带表示在图像参数集( pic_parameter— set— rbsp, PPS )中的图像初 始量化减去 26 ( pic_init_qp_minus26, pic_init_qp_minus26是 H.264图像参 数集中的一个语法元素)所在的传输包不丟失情况下, 条带的第一个传输包 至用于获取所述条带的量化参数条带的量化参数差量 ( slice— qp— delta , slice— qp— delta是 H.264条带头中的一个语法元素)所在的传输包都不发生丟 失; 换句话说, 条带的第一个比特至用于解析出条带的量化参数差量 ( slice— qp— delta, slice— qp— delta是 H.264条带头中的一个语法元素) 的比特 都不发生丟失, 条带的量化参数差量用于获取所述条带的量化参数; 这样, 可以正确解析出该条带的量化参数。 其中, 条带的量化参数的计算公式为: i— slice— qp = 26 + pic_init_qp_minus26十 slice— qp— delta。 N为正整数, Nl为正 整数且 N1≤N。 进一步地, 可以将该 N1个条带的量化参数的平均值确定为 用于预测编码质量的上述视频的量化参数, 如下列 (1 ) 式视频的量化参数 为 f— video— qp:
N
^ i _ slice _ qpt
f video qp =— ( 1 )
_ N\
上式中, i— slice _ 9 1为 1^1个条带的量化参数中第 i个条带的量化参 数。
可选地,上述 N1还可以是不发生丟包的条带的条带的量化参数的数目。 可选地, 还可以将该 N1个条带的量化参数的加权平均值确定为用于预 测编码质量的上述视频的量化参数, w;为该第 i个条带所对应的权重, 可以 根据需要对不同的条带赋予不同的权重。 如下列公式所示:
N 1
wi · i _ slice _ qpt
f _ video _ qp= ^ l i=\
进一步地,还可以对上述各式得到的视频的量化参数 f— video— qp进行归 一化调整, 将调整后的视频的量化参数用于预测编码质量。 应理解, 上述例子仅仅是示例性的, 本发明实施例对如何获得用于预测 编码质量的视频的量化参数不作限制,也可以将视频中所有正确解析的帧内 编码帧和帧间编码帧的条带的量化参数的平均值或加权平均值确定为用于 预测编码质量的上述视频的量化参数,还可以将视频中所有条带和宏块的量 化参数的平均值或加权平均值确定为用于预测编码质量的视频的量化参数。 等等。
可选地, 在视频帧的传输过程中, 还可能会产生丟包, 可以选择丟包或 没有丟包的条带进行计算, 应理解, 本发明实施例对此并不限定。
可选地, 上述视频帧可以是帧间编码帧和 /或帧内编码帧, 其中, 帧间编 码帧可以是作为参考帧的帧间编码帧(也称为帧间参考帧 ),如 P帧或 B帧; 也可以是不作为参考帧的帧间编码帧,如 b帧;帧内编码帧可以是 I帧或 IDR ( Instantaneous Decoding Refresh, 解码立即刷新) 帧, 也称为帧内参考帧。 等等。 应理解, 本发明实施例对此并不限定。
优选地,根据视频流的视频帧的条带的像素字节数确定视频的内容复杂 度, 首先根据视频流的视频帧的条带的像素字节数确定视频帧的内容复杂 度,再根据视频帧的内容复杂度确定视频的内容复杂度,具体地,如下列(2 ) - ( 4 ) 式:
视频帧的 Num个条带中的第 k个条带的像素字节数为:
r Ί. , . , i slice size,
j _ slice _ byte _ per _ pixelk =—— ― ― ( 2 )
― — — ― i—slice—pixelk
其中, i— slice— sizek为所述视频帧的 Num个条带中第 k个条带的字节数, i— slice_pixelk为所述视频帧的 Num个条带中第 k个条带的像素数, k为正整 数, 在所述视频帧中, k的取值从 1到 Num,Num为正整数。
需要说明的是, 本发明实施例中对像素字节数还可以做一些变形, 例如 可以变形为像素比特数, 然后采用像素比特数进行相应的操作, 该种变形并 未脱离本发明实施例的实质, 仍应属于本发明的保护范围。
视频帧的 Num个条带中的第 k个条带的内容复杂度为:
f _ slice _ content _ complexity k =
Figure imgf000012_0001
a i _ slice _gpk] - f _ slice _ byte _ per _ pixel k + b\i_ slice _ qpl
其中, a[i— slice— qpk]和 b[i— slice— qpk]均为与所述视频帧的 Num个条带中 k个条带的量化参数对应的值。 视频帧的内容复杂度为:
Num
∑ / _ slice _ content _ complexity f _ frame _ content _ complexity =
Num
( 4 ) 可选地, Num可以为视频帧的所有条带的数目(包括丟包和不丟包的条 带数目 ), Num也可以为视频帧的不丟包的条带的数目或者丟包的条带的数 本发明实施例对此不限定。
进一步地, 根据视频流的视频帧的内容复杂度确定视频的内容复杂度, 可以将多个视频帧的内容复杂度的平均值或加权平均值确定为用于预测编 码质量的上述视频的内容复杂度, 具体地, 如下列 (5 ) 式所示:
视频帧的内容复杂度为:
M
_ frame _ content _ complexity j f _ video _ content _ complexity =
M
( 5 ) 其中, 视频流包括 M个视频帧, M为正整数, j为正整数且取值从 1到 M。 优选地, M个视频帧均不丟包。
视频的内容复杂程度可以包括时间方向和 /或空间方向,在某一分辨率的 某一固定的量化参数下,视频的内容复杂度越大,失真越不容易被人眼发现。 通过上述技术方案得到的视频内容复杂度较符合视频内容复杂度越大, 失真 越不容易被人眼发现的人眼主观感受,从而进一步地提高了预测编码质量的 准确性。
例如, 上述(2 ) - ( 5 ) 式中的视频帧为帧内参考帧, 即视频的内容复 杂度可通过帧内参考帧的条带像素字节数确定(使用至少一个帧内参考帧), 在该例子中, 由帧内参考帧的条带像素字节数确定的视频的内容复杂度能够 有效地体现出视频的空间复杂度 f— video— content— complexity— sec , 可选地, 式子中的 a[i— slice— qpk]和 b[i— slice— qpk]可以通过训练得到, 不同的视频分辨 率和 /或编码格式和 /或视频采集格式(即: 隔行扫描方式和逐行扫描方式) 可以使用相同的值或者不同的值。 以 Η.264中具有 52个量化参数 (分别是 0 至 51 ) 为例, 相应地, 通过训练得到的在不同分辨率下的 a[i— slice— qpk]和 slice qpk] , 分别与量化参数值 0至 51相对应, 示例性地表示如下: a[i— slice— qpk] ,
标清分辨率:
{24.78954, 24.78954, 25.23854, 25.51193, 25.74990, 25.97533, 26.19479 26.28303, 26.49158, 26.56645, 26.53197, 26.62563, 26.69239, 26.65409,
26.79309, 26.80578, 26.84816, 27.08741, 27.25370, 27.36097, 27.56078,
27.70162, 27.85621, 28.04059, 28.17621, 28.23445, 28.41471, 28.45078,
28.54265, 28.60014, 28.62930, 28.64529, 28.74102, 28.75523, 28.76358,
28.74681, 28.77488, 28.73642, 28.79531, 28.69430, 28.72766, 28.60666, 28.49484, 28.35642, 28.07614, 27.90134, 27.57123, 27.01405, 26.65987,
26.31439, 25.52575, 25.01169}
1280x720分辨率:
{16.17209, 17.45819, 17.80732, 18.02041, 18.18083, 18.52479, 19.03342
19.06581, 19.41564, 19.85189, 20.07956, 20.81183, 21.43127, 21.83287,
22.61658, 23.14807, 23.92571, 25.20184, 26.03683, 26.68701, 27.49974,
28.12203, 28.66205, 29.27020, 29.69070, 29.92960, 30.40275, 30.60385,
30.85636, 31.06785, 31.26051, 31.35589, 31.63646, 31.76881, 31.92259,
32.08798, 32.28134, 32.36179, 32.60119, 32.61653, 32.75291, 32.73418,
32.72940, 32.70158, 32.59009, 32.41000, 32.21505, 31.76353, 31.23468,
30.87401, 30.01071, 29.31316}
全高清分辨率( 1920x1080 ):
{15.75673, 16.17239, 17.33657, 18.09218, 18.78856, 19.85244, 20.94081
21.42377, 25.25608, 25.36929, 25.37671, 25.59413, 25.77414, 25.89431,
26.16539, 26.37098, 26.71202, 27.45373, 27.99336, 28.43923, 29.01115,
29.49924, 29.89337, 30.32379, 30.59313, 30.74944, 31.01314, 31.10389,
31.21737, 31.28295, 31.38585, 31.36863, 31.44693, 31.40169, 31.43938,
31.39075, 31.36072, 31.33672, 31.26816, 31.16160, 31.03165, 30.80631,
30.57609, 30.36353, 30.06076, 29.62381, 29.37353, 29.05716, 28.60942,
28.52338, 28.40104, 28.52280}
b[i— slice— qp ,
标清分辨率: {13.39250, 13.39250, 13.97091, 14.53803, 15.25528, 16.13630, 16.99497, 17.66163, 18.80068, 19.89785, 21.20091, 22.86877, 24.44105, 25.98037, 28.04957, 30.07985, 32.07935, 34.30203, 36.32256, 38.18652, 40.93258, 43.77054, 46.53546, 50.53632, 54.36178, 57.82423, 63.29899, 69.18878, 75.07466, 83.80263, 91.47496, 99.18949, 111.47580, 124.34650, 136.49900, 156.17670, 176.23080, 192.16970, 223.83720, 251.77270: 285.92790, 333.53770, 388.41820, 435.09860, 531.05070, 633.24080, 760.16820, 948.15240, 1168.53720, 1361.84570, 1759.43160, 2040.35460}
1280x720分辨率:
{33.81798, 33.05324, 35.11725, 36.95499, 39.10951, 41.62373, 43.87256,
45.95354, 49.32386, 51.87803, 54.92251, 58.42482, 61.62755, 64.56505, 69.19412, 73.35919, 76.10406, 78.96517, 81.95586, 84.59924, 89.05335, 93.59975, 98.31476, 105.41810, 112.34964, 118.73374, 129.00992, 140.01562, 151.12381, 167.62430, 182.02425, 196.08347, 218.72591, 241.16108, 263.35157: 295.99927, 329.06899, 355.66280, 407.64235, 452.09915, 508.72302, 585.36672: 671.43978, 741.49561, 891.18944, 1051.86892, 1246.04333, 1527.50615, 1894.63282, 2204.87735, 2879.95903, 3390.89788}
全高清分辨率( 1920x1080 ):
{25.92973, 26.42403, 26.72231, 27.10874, 27.55908, 27.59167, 27.40409, 27.63129, 21.08740, 22.32786, 23.78112, 25.55635, 27.25511, 28.80079, 31.33600, 33.71534, 35.51380, 37.14249, 38.57997, 39.75292, 41.50986, 43.25411, 45.08496, 47.92251, 50.97660, 53.82247, 58.50549, 64.00109, 69.59487, 78.31654, 84.35147, 92.89916, 105.12040, 119.83478, 131.13182, 152.46046, 175.28796, 191.40711, 231.17849, 262.14953: 311.33306, 374.98524, 454.98602, 524.68907, 656.91124, 830.55605, 990.09180, 1196.94617, 1493.32352, 1667.34794, 1966.34090, 2099.62991 }
又例如, 上述(2 ) - ( 5 ) 式中的视频帧为帧间参考帧, 即视频的内容 复杂度可通过帧间参考帧的条带像素字节数确定(使用至少一个帧间参考 帧), 在该例子中, 由帧间参考帧的条带像素字节数确定的视频的内容复杂 度能够有效地体现出视频的时间复杂度 f— video— content— complexity— tec , 可 选地, 式子中的 a[i— slice— qpk]和 b[i— slice— qpk]可以通过训练得到, 不同的视 频分辨率和 /或编码格式和 /或视频采集格式(即: 隔行扫描方式和逐行扫描 方式)可以使用相同的值或者不同的值。 以 H.264中具有 52个量化参数 (分 别是 0至 51 )为例,相应地,通过训练得到的在不同分辨率下的 a[i— slice— qpk] 和 b[i— slice— qpk] , 分别与量化参数值 0至 51相对应。
应理解, 上述例子中视频帧的选取, 条带的量化参数的数目, 以及 a[i— slice— qpk]和 b[i— slice— qpk]的值及得到的方式等等仅仅是示例性的, 而非 要限制本发明的范围。
实验中发现, 当视频中有编码压缩失真时, 视频编码质量会随着 QP变 大而降低。 可选地, 基于上述实验, 在步骤 102 中, 可以根据上述(5 ) 式 的视频的内容复杂度 f— video— content— complexity和上述( 1 ) 式的视频的量
4匕参数 f video qp预则视频编码质量 d— compression— quality— value , ^口下歹 ll
( 6 ) 式:
预测视频编码质量为:
d _ compression _ quality _ value = ai ( 6 )
Figure imgf000016_0001
α4 - α5 · f _ video _ content _ complexity
其中, a5和 a6均为常数。
可选地, 在不同的视频分辨率和 /或编码格式和 /或视频采集格式(即: 隔行扫描方式和逐行扫描方式)下, 至&6的取值可以相同或不同。 在标清 分辨率下 1至 a6的取值与在全高清分辨率 ^至 a6的取值可以是相同或不同 的, 应理解, 本发明实施例对此不作限定。
当视频编码质量 d— compression— quality— value的值越大表示视频质量越 好, 而当视频编码质量 d— compression— quality— value的值越小表示视频质量 越差。
可选地, a2取值可以为视频编码质量范围的最大值与最小值之差。例如, 在 5分制的编码质量的度量中, 编码质量最大值为 5 , 最小值为 1 , 因此 a2 取值为 4。 又例如, a2取值可以为通过训练或经验得到的视频编码质量范围 的最大值与最小值之差, 如通过训练或经验得到编码质量最大值为 4.5 , 以 及编码质量最小值为 1.5 , 因此 a2取值为 3。 应理解, 本发明实施例预测编 码质量的方式可以应用到其它分制的编码质量的度量中,本发明实施例对此 不作限制。
可选地, a取值可以为视频编码质量范围的最小值。
可选地, a5和 a6的取值还可以根据不同的情况进行取值, 应理解, 本发明实施例对此不作限定。 例如, ai, a2, a3, a4, a5和 a6的取值 还可以根据不同的情况进行训练或经验取值。
可选地, a5和 a6的取值示例性地表示如下:
标清:
Figure imgf000017_0001
a2 = 2.9116,
a3= 1.0,
a4 = 41.5,
a5 = 4.7,
a6= 13.0
1280x720:
ai = 1.0519,
a2 = 3.3876,
a3= 1.0,
a4 = 40.0,
a5 = 0.75,
a6= 10.0
1920x1080, 隔行扫描(1920xl080i):
ai = 1.2294,
a2 = 3.1092,
a3= 1.0,
a4 = 41.5,
a5 = 0.65,
a6= 10.5
1920x1080, 逐行扫描(1920x1080p): ai = 1.2294,
a2 = 3.1092,
a3 = 1.0,
a4 = 43.0,
a5 =0.85 ,
a6 = 12.0
需要指出的是, 以其它等效的公式表示上述编码质量的方式都落入本发 明的范围, 例如, 视频编码失真质量 d— compression— artifact— value由视频编 码质量的最大值减去上述视频编码质量获得的, 如下式(7 )所示:
d compression artifact value =
Figure imgf000018_0001
此时, 当视频编码失真质量越小, 视频质量越好, 而当视频编码失真质 量越大, 视频质量越差。
可选地, 可以对上述(5 )式至(7 )式中的视频的内容复杂度进行调整, 将调整后的 f— video— content— complexity— n作为所述视频的内容复杂度,调整 方式如下列 (8 )所示:
f _ video _ content _ complexity _ n =
其中, s和 numl为常数。例如,进行归一化调整时, s取值为 1.0, numl 取值为 60.0 , f—vidio— content— complexity— n取值范围为 [0.0, 1.0]。
应理解, 视频的内容复杂度可以是时间复杂度(如视频帧只考虑帧间参 考†贞 ) , 表示为 f— video— content— complexity=f— video— content— complexity— tec ( 上 述 ( 5 ) 式 ) 或 f— video— content— complexity— n=f— video— content— complexity— ntcc (上述 ( 8 )式, 进行了调整的时间复杂度); 也可以是空间复杂度(如视频帧只考虑帧内参 考†贞 ) , 表示为 f— video— content— complexity=f— video— content— complexity— sec ( 上 述 ( 5 ) 式 ) 或 f— video— content— complexity— n=f— video— content— complexity— nscc (上述 ( 8 ) 式, 进行了调整的空间复杂度); 还可以是结合时间复杂度和空间复杂度, 表示为 f— video— content— complexity=fimcl ( f— video— content— complexity— tec , f— video— content— complexity— sec ) 或 f— video— content— complexity— n=fimc2
( f— video— content— complexity— ntcc, f— video— content— complexity— nscc ), 本发 明实施例对此并不限定。
需要指出的是, 在采用 (6 )式或(7 )式预测编码质量或视频编码失真 质量的实施例中, 可以采用经过(8 )式进行调整的视频的内容复杂度, 即: 预测视频编码质量可以为:
d _ compression _ quality _ value = a、 ( 9 )
Figure imgf000019_0001
αΛ - α5 · _ vi eo _ content _ comp exity _ n 其中, a5和 a6均为常数。
预测视频编码质量也可以为:
d _ compression _ artifact _ value =
f video qp
a2
a -a5 - f video content complexity _n
( 10 ) α, Η f _video_qp
a -a5 · f _ video _ content _ complexity _ n
应理解,调整后的视频的内容复杂度 f— video— content— complexity— n可以 是进行归一化调整后的视频的时间复杂度 f— video— content— complexity— ntcc , 或 者 是 进 行 归 一 化 调 整 后 的 视 频 的 空 间 复 杂 度 f— video— content— complexity— nscc , 还 可 以 是 时 间 复 杂 度 f— video— content— complexity— ntcc 和 空 间 复 杂 度 f— video— content— complexity— nscc的组合。 应理解, 本发明实施例对此不作限 定。
例如, 采用时间复杂度 f— video— content— complexity— ntcc 和空间复杂度 f— video— content— complexity— nscc的组合和视频的量 4匕参数 f video qp预测视 频编码质量 d— compression— quality— value, 即在(6 ) 式中视频的内容的复杂 度采用调整后的视频的内容的复杂度 f— video— content— complexity— n, 可将 a5*f— video— content— complexity— n 替 换 成 a51 *f— video— content— complexity— ntcc+a52*f— video— content— complexity— nscc ,其 中, a51和 a52可以相同或不同。
通过本发明实施例预测编码质量的方式更符合人眼主观感受,从而进一 步地提高了预测编码质量的准确性。
当然, 视频的内容复杂度可以根据其它视频帧的编码信息来确定, 如视 频帧的条带的编码字节数、运动矢量差值、运动矢量值、宏块编码模式、 DCT ( Discrete Cosine Transform, 离散余弦变换) 系数等。 例如, 视频的内容复 杂度表示为所有帧内编码帧 (I/IDR 帧) 的离散余弦变换 DCT 系数的 AC ( alternating Current, 交流 )系数的平均值,或者所有帧内编码帧( I/IDR帧 ) 的离散余弦变换 DCT系数的 DC ( Direct Current, 直流分量)系数的平均值, 或者所有帧内编码帧( I/IDR帧 )的离散余弦变换 DCT系数的平均值, 或者 帧间编码帧的离散余弦变换 DCT系数的 AC系数的平均值, 或者帧间编码 帧的离散余弦变换 DCT系数的 DC系数平均值, 或者帧间编码帧的离散余 弦变换 DCT 系数的平均值, 或者帧内编码帧和帧间编码帧的离散余弦变换 DCT 系数的平均值。 又例如, 视频的内容复杂度通过宏块编码模式 (如 16x16,16x8,8x16等等)来描述,宏块编码模式越多,视频的内容复杂度越高。 再例如, 视频的内容复杂度通过运动矢量来描述, 运动矢量越大, 视频的内 容复杂度越高。 再例如, 视频的内容复杂度通过帧间编码块与帧内编码块的 比例来表示, 比例越大, 视频的内容复杂度越高。 等等。 应理解, 本发明实 施例对此并不限定。
图 2是本发明一个实施例的视频编码质量的评估设备的结构框图。视频 编码质量的评估设备 200包括获取单元 201、确定单元 202和预测单元 203。
获取单元 201 , 用于获取视频流的视频帧的条带的量化参数, 和视频流 的视频帧的条带的像素字节数。
确定单元 202, 用于根据获取单元 201获取的视频流的视频帧的条带的 量化参数确定视频的量化参数, 并根据获取单元 201获取的视频流的视频帧 的条带的像素字节数确定视频的内容复杂度。
预测单元 203 , 用于根据确定单元 202确定的视频的内容复杂度和视频 的量化参数预测视频编码质量。
在本发明实施例中,量化参数也可以采用另一种表达方式,量化步长( QP Step ) 来进行预测视频编码质量。 为了描述方便, 本发明实施例预测视频编 码质量的模型中将以量化参数来举例说明, 应理解, 以 QP Step代替 QP的 表达方式也落入本发明范围内。
本发明实施例通过根据获取的视频流的视频帧的条带的量化参数确定 视频的量化参数, 并根据获取的视频流的视频帧的条带的像素字节数确定视 频的内容复杂度,通过视频的内容复杂度和视频的量化参数预测视频编码质 量。 因此, 通过考虑了视频内容特性得到的模型所预测出的编码质量更符合 人眼主观感受, 从而提高预测的准确性。
设备 200可实现图 1 的方法中涉及视频编码质量的评估设备的各个步 骤, 为避免重复, 不再详细描述。
可选地, 作为一个实施例, 确定单元 202可以具体用于: 将视频流的视 频帧的条带的量化参数的平均值或加权平均值确定为视频的量化参数。具体 地, 确定单元 202可以用于:
确定视频的量化参数为 f— video— qp:
N
^ ί _ slice _ qpt
f video qp=— , 或者
N \
^ wt · i _ slice _ qpt
f _ video _ qp = ^ i=l
其中, Nl为视频流的条带的量化参数的数目, i— slicc— qp^ N1个 条带的量化参数中第 i个条带的量化参数, N1为正整数, w;为该第 i个条带 所对应的权重。
优选地, 该 N1为正确解析出的条带的量化参数的总数, 正确解析出的 条带的量化参数表示在条带的第一个传输包至用于获取条带的量化参数的 条带量化参数差量所在的传输包都不丟失的情况下,解析出的条带的量化参 数为正确的。
进一步地, 确定单元 202 还可以对上式得到的视频的量化参数 f— video— qp进行归一化调整,将调整后的视频的量化参数用于预测编码质量。
可选地, 确定单元 202可以具体用于:
确定视频帧的 Num个条带中的第 k个条带的像素字节数为: i slice size.
f _ slice _ byte _ per _ pixel}
i—slice—pixelk
其中, i— slice— sizek为视频帧的 Num个条带中第 k个条带的字节数, i— slice_pixelk为视频帧的 Num个条带中第 k个条带的像素数, k为正整数, 在视频帧中, k的取值从 1到 Num,Num为正整数;
具体用于: 确定视频帧的 Num个条带中的第 k个条带的内容复杂度为: f _ slice _ content _ complexity k
(i i _ slice _qpk ' f— slice _ byte _ per _ pixelk + b i_ slice _ qpk
其中, a[i— slice— qpk]和 b[i— slice— qpk]均为与视频帧的 Num个条带中的第 k个条带的量化参数对应的值;
具体用于: 确定视频帧的内容复杂度为:
Num
∑ / _ slice _ content _ complexity k=l
f _ frame _ content _ complexity
Num
具体用于: 根据视频帧的内容复杂度确定视频的内容复杂度。
可选地, 上述视频帧可以是帧间编码帧和 /或帧内编码帧, 其中, 帧间编 码帧可以是作为参考帧的帧间编码帧(也称为帧间参考帧 ),如 P帧或 B帧; 也可以是不作为参考帧的帧间编码帧,如 b帧;帧内编码帧可以是 I帧或 IDR ( Instantaneous Decoding Refresh, 解码立即刷新) 帧, 也称为帧内参考帧。 等等。 应理解, 本发明实施例对此并不限定。
进一步地, 确定单元可以具体用于: 确定视频的内容复杂度为:
M
_ frame _ content _ complexity j f _ video _ content _ complexity =
M
其中, 视频流包括 M个视频帧, M为正整数, j为正整数且取值从 1到 M。
可选地, Num可以为视频帧的所有条带的数目(包括丟包和不丟包的条 带数目 ), Num也可以为视频帧的不丟包的条带的数目或者丟包的条带的数 本发明实施例对此不限定。
可选地, M个视频帧均没有丟包。
可选地, 视频帧可以包括帧内编码帧和 /或帧间编码帧。 视频的内容复杂程度可以包括时间方向和 /或空间方向,在某一分辨率的 某一固定的量化参数下,视频的内容复杂度越大,失真越不容易被人眼发现。 通过上述技术方案得到的视频内容复杂度较符合视频内容复杂度越大, 失真 越不容易被人眼发现的人眼主观感受,从而进一步地提高了预测编码质量的 准确性。
可选地, 预测单元 203可以具体用于: 预测视频编码质量为:
d _ compression _ quality _ value =
a. α - α5 · f _ video _ content _ complexity
其中, a5和 a6均为常数。
当视频编码质量 d— compression— quality— value的值越大表示视频质量越 好, 而当视频编码质量 d— compression— quality— value的值越小表示视频质量 越差。
可选地, a2取值可以为视频编码质量范围的最大值与最小值之差。例如, 在 5分制的编码质量的度量中, 编码质量最大值为 5 , 最小值为 1 , 因此 a2 取值为 4。 又例如, a2取值可以为通过训练或经验得到的视频编码质量范围 的最大值与最小值之差, 如通过训练或经验得到编码质量最大值为 4.5 , 以 及编码质量最小值为 1.5 , 因此 a2取值为 3。 应理解, 本发明实施例预测编 码质量的方式可以应用到其它分制的编码质量的度量中,本发明实施例对此 不作限制。
可选地, a取值可以为视频编码质量范围的最小值。
可选地, a5和 a6的取值还可以根据不同的情况进行取值, 应理解, 本发明实施例对此不作限定。 例如, ai , a2, a3, a4, a5和 a6的取值 还可以根据不同的情况进行训练或经验取值。
需要指出的是, 以其它等效的公式表示上述编码质量的方式都落入本发 明的范围, 例如, 视频编码失真质量 d— compression— artifact— value由视频编 码质量的最大值减去上述视频编码质量获得的, 如上述(7 ) 式, 此时, 当 视频编码失真质量越小, 视频质量越好, 而当视频编码失真质量越大, 视频 质量越差。
可选地, 确定单元 202可以具体用于将视频的内容复杂度调整为: video content complexity
Figure imgf000024_0001
其中, s和 numl为常数。例如,进行归一化调整时, s取值为 1.0, numl 取值为 60.0 , f—vidio— content— complexity— n取值范围为 [0.0, 1.0]。
应理解, 视频的内容复杂度可以是时间复杂度(如视频帧只考虑帧间参 考†贞 ) , 表示为 f— video— content— complexity=f— video— content— complexity— tec ( 上 述 ( 5 ) 式 ) 或 f— video— content— complexity— n=f— video— content— complexity— ntcc (上述 ( 8 )式, 进行了调整的时间复杂度); 也可以是空间复杂度(如视频帧只考虑帧内参 考†贞 ) , 表示为 f— video— content— complexity=f— video— content— complexity— sec ( 上 述 ( 5 ) 式 ) 或 f— video— content— complexity— n=f— video— content— complexity— nscc (上述 ( 8 ) 式, 进行了调整的空间复杂度); 还可以是结合时间复杂度和空间复杂度, 表示为 f— video— content— complexity=fimcl ( f— video— content— complexity— tec , f— video— content— complexity— sec ) 或 f— video— content— complexity— n=fimc2 ( f— video— content— complexity— ntcc , f— video— content— complexity— nscc ), 本发 明实施例对此并不限定。
需要指出的是, 在采用 (6 )式或(7 )式预测编码质量或视频编码失真 质量的实施例中, 可以采用经过(8 ) 式进行调整的视频的内容复杂度, 调 整后的视频的内容复杂度 f— video— content— complexity— n可以是进行归一化调 整后的视频的时间复杂度 f— video— content— complexity— ntcc, 或者是进行归一 化调整后的视频的空间复杂度 f— video— content— complexity— nscc,还可以是时 间 复 杂 度 f— video— content— complexity— ntcc 和 空 间 复 杂 度 f— video— content— complexity— nscc的组合。 应理解, 本发明实施例对此不作限 定。
例如, 采用时间复杂度 f— video— content— complexity— ntcc 和空间复杂度 f— video— content— complexity— nscc的组合和视频的量 4匕参数 f video qp预测视 频编码质量 d— compression— quality— value, 即在(6 ) 式中视频的内容的复杂 度采用调整后的视频的内容的复杂度 f— video— content— complexity— n, 可将 a5*f— video— content— complexity— n 替 换 成 a51 *f— video— content— complexity— ntcc+a52*f— video— content— complexity— nscc ,其 中, a51和 a52可以相同或不同。
通过本发明实施例预测编码质量的方式更符合人眼主观感受,从而进一 步地提高了预测编码质量的准确性。
图 3是本发明另一个实施例的视频编码质量的评估设备的结构框图。 如图 3所示, 一般包括至少一个处理器 410, 例如 CPU, 至少一个端口 420, 存储器 430, 和至少一个通信总线 440。 通信总线 440用于实现这些装 例如计算机程序; 可选地, 设备 300可包括用户接口 450 , 用户接口 450包 括但不限于显示器, 键盘和点击设备, 例如鼠标、 轨迹球( trackball )、 触感 板或者触感显示屏。 存储器 430可能包含高速 RAM存储器, 也可能还包括 非易失性存储器( non-volatile memory ), 例如至少一个磁盘存储器。
在一些实施方式中, 存储器 430存储了如下的元素, 可执行模块或者数 据结构, 或者他们的子集, 或者他们的扩展集。
操作系统 432, 包含各种系统程序, 用于实现各种基础业务以及处理基 于硬件的任务。
应用模块 434, 包含各种应用程序, 用于实现各种应用业务。
应用模块 434中包括但不限于获取单元 201、 确定单元 202和预测单元
203。
应用模块 434中各单元的具体实现参见图 2所示实施例中的相应单元, 在此不赘述。
上述视频编码质量的评估设备可以是终端, 例如, 可以是便携式、 袖珍 式、 手持式、 计算机内置的或者车载的移动装置, 或者, 设备还可以是服务 器。 等等。
本领域普通技术人员可以意识到, 结合本文中所公开的实施例描述的各 示例的单元及算法步骤, 能够以电子硬件、 或者计算机软件和电子硬件的结 合来实现。 这些功能究竟以硬件还是软件方式来执行, 取决于技术方案的特 定应用和设计约束条件。 专业技术人员可以对每个特定的应用来使用不同方 法来实现所描述的功能, 但是这种实现不应认为超出本发明的范围。
所属领域的技术人员可以清楚地了解到, 为描述的方便和筒洁, 上述描 述的系统、 装置和单元的具体工作过程, 可以参考前述方法实施例中的对应 过程, 在此不再赘述。
在本申请所提供的几个实施例中, 应该理解到, 所揭露的系统、 装置和 方法, 可以通过其它的方式实现。 例如, 以上所描述的装置实施例仅仅是示 意性的, 例如, 所述单元的划分, 仅仅为一种逻辑功能划分, 实际实现时可 以有另外的划分方式,例如多个单元或组件可以结合或者可以集成到另一个 系统, 或一些特征可以忽略, 或不执行。 另一点, 所显示或讨论的相互之间 的耦合或直接耦合或通信连接可以是通过一些接口, 装置或单元的间接耦合 或通信连接, 可以是电性, 机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的,作 为单元显示的部件可以是或者也可以不是物理单元, 即可以位于一个地方, 或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或 者全部单元来实现本实施例方案的目的。
另外, 在本发明各个实施例中的各功能单元可以集成在一个处理单元 中, 也可以是各个单元单独物理存在, 也可以两个或两个以上单元集成在一 个单元中。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使 用时, 可以存储在一个计算机可读取存储介质中。 基于这样的理解, 本发明 的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部 分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质 中, 包括若干指令用以使得一台计算机设备(可以是个人计算机, 服务器, 或者网络设备等)执行本发明各个实施例所述方法的全部或部分步骤。 而前 述的存储介质包括: U盘、移动硬盘、只读存储器( ROM, Read-Only Memory )、 随机存取存储器(RAM, Random Access Memory ), 磁碟或者光盘等各种可 以存储程序代码的介质。

Claims

权利要求
1、 一种视频编码质量的评估方法, 其特征在于, 包括:
获取视频流的视频帧的条带的量化参数和所述视频流的视频帧的条带 的像素字节数;
根据所述视频流的视频帧的条带的量化参数确定视频的量化参数,并根 据所述视频流的视频帧的条带的像素字节数确定视频的内容复杂度;
根据所述视频的内容复杂度和所述视频的量化参数预测视频编码质量。
2、 如权利要求 1所述的方法, 其特征在于, 所述根据视频流的视频帧 的条带的量化参数确定视频的量化参数, 包括:
将所述视频流的视频帧的条带的量化参数的平均值或加权平均值确定 为所述视频的量化参数。
3、 如权利要求 2所述的方法, 其特征在于, 所述将所述视频流的视频 帧的条带的量化参数的平均值或加权平均值确定为所述视频的量化参数, 包 括:
^ ί _ slice _ qpt
f video qp=— , 或者
N \
^ wt · i _ slice _ qpt
f _ video _ qp = ^ i=l 其中, f— video— qp为所述视频的量化参数, N1为所述视频流的条带的量 化参数的数目, 为所述 N1个条带的量化参数中第 i个条带的 量化参数, N1为正整数, w;为该第 i个条带所对应的权重。
4、 如权利要求 1-3任一项所述的方法, 其特征在于, 所述根据所述视 频流的视频帧的条带的像素字节数确定视频的内容复杂度, 包括:
确定视频帧的 Num个条带中的第 k个条带的像素字节数为:
r Ί. , . , i slice size,
j _ slice _ byte _ per _ pixelk = = ―
― — — ― i—slice _pixelk
其中, i— slice— sizek为所述视频帧的 Num个条带中第 k个条带的字节数, i— slice_pixelk为所述视频帧的 Num个条带中第 k个条带的像素数, k为正整 在所述视频帧中, k的取值从 1到 Num,Num为正整数; 确定所述视频帧的 Num个条带中的第 k个条带的内容复杂度为 f _ slice _ content _ complexity k
(i i _ slice _qpk ' f— slice _ byte _ per _ pixelk +b i_ slice _ qpk 其中, a[i— slice— qpk]和 b[i— slice— qpk]均为与所述视频帧的 Num个条带中 % k个条带的量化参数对应的值;
确定所述视频帧的内容复杂度为:
Num
∑ / _ slice _ content _ complexity k f _ frame _ content _ complexity =―
Num
根据所述视频帧的内容复杂度确定所述视频的内容复杂度。
5、 如权利要求 4所述的方法, 其特征在于, 所述根据所述视频帧的内 容复杂度确定所述视频的内容复杂度, 包括:
M
_ frame _ content _ complexity j f _ video _ content _ complexity =
M
其中, f— video— content— complexity为所述视频的内容复杂度, 所述视频 流包括 M个视频帧, M为正整数, j为正整数且取值从 1到 M。
6、 如权利要求 5所述的方法, 其特征在于, 所述根据所述视频的内容 复杂度和所述视频的量化参数预测视频编码质量, 包括:
d _ compression _ quality _ value =
a2
1
Figure imgf000028_0001
α4 - 5 · f _ video _ content _ complexity
其中, d— compression— quality— value 为所述视频编码质量, a a2, a3, a4, a5和 a6均为常数。
7、 如权利要求 6所述的方法, 其特征在于, 所述 a2取值为:
视频编码质量范围的最大值与最小值之差。
8、 如权利要求 6或 7所述的方法, 其特征在于, 所述 1取值为: 视频编码质量范围的最小值。
9、 如权利要求 4-8任一项所述的方法, 其特征在于, 还包括, 对所述 视频的内容复杂度进行调整: video content complexity
Figure imgf000029_0001
将调整后的 f— video— content— complexity— n作为所述视频的内容复杂度, s和 numl均为常数。
10、 如权利要求 9所述的方法, 其特征在于, 所述 s取值为 1.0, 所述 numl取值为 60.0, 所述 f—vidio— content— complexity— n取值范围为 [0.0, 1.0]。
11、 如权利要求 5-10任一项所述的方法, 其特征在于, 所述 M个视频 帧均没有丟包。
12、 如权利要求 5-11 任一项所述的方法, 其特征在于, 所述视频帧包 括帧内编码帧和 /或帧间编码帧。
13、 如权利要求 4-12任一项所述的方法, 其特征在于, Num为所述视 频帧中条带的总数。
14、 如权利要求 3-13任一项所述的方法, 其特征在于, 所述 N1为正确 解析出的条带的量化参数的数目, 所述正确解析出的条带的量化参数表示在 条带的第一个传输包至用于获取所述条带的量化参数的条带量化参数差量 所在的传输包都不丟失的情况下, 解析出的条带的量化参数为正确的。
15、 一种视频编码质量的评估设备, 其特征在于, 包括:
获取单元,用于获取视频流的视频帧的条带的量化参数和所述视频流的 视频帧的条带的像素字节数;
确定单元,用于根据所述获取单元获取的所述视频流的视频帧的条带的 量化参数确定视频的量化参数, 并根据所述获取单元获取的所述视频流的视 频帧的条带的像素字节数确定视频的内容复杂度;
预测单元,用于根据所述确定单元确定的所述视频的内容复杂度和所述 视频的量化参数预测视频编码质量。
16、 如权利要求 15所述的设备, 其特征在于,
所述确定单元具体用于: 将所述视频流的视频帧的条带的量化参数的平 均值或加权平均值确定为所述视频的量化参数。
17、 如权利要求 16所述的设备, 其特征在于,
所述确定单元具体用于: 确定所述视频的量化参数为: N l
^ ί _ slice _ qpt
f — video _ qp= ^ ― , 或者
Ν\
N l
^ wt · i _ slice _ qpt
f _ video _ qp = ^
Figure imgf000030_0001
其中, f— video— qp为所述视频的量化参数, Nl为所述视频流的条带的量 化参数的数目, 为所述 N1个条带的量化参数中第 i个条带的 量化参数, N1为正整数, w;为该第 i个条带所对应的权重。
18、 如权利要求 15-17任一项所述的设备, 其特征在于,
所述确定单元具体用于:
确定视频帧的 Num个条带中的第 k个条带的像素字节数为:
i slice size,
f _ slice _ byte _ per _ pixel}
i—slice—pixelk
其中, i— slice— sizek为所述视频帧的 Num个条带中第 k个条带的字节数, i— slice_pixelk为所述视频帧的 Num个条带中第 k个条带的像素数, k为正整 数, 在所述视频帧中, k的取值从 1到 Num,Num为正整数;
具体用于: 确定所述视频帧的 Num个条带中的第 k个条带的内容复杂 度为:
f _ slice _ content _ complexity k
(i i _ slice _qpk ' f— slice _ byte _ per _ pixel k + b i_ slice _ qpk
其中, a[i— slice— qpk]和 b[i— slice— qpk]均为与所述视频帧的 Num个条带中 的第 k个条带的量化参数对应的值;
具体用于: 确定所述视频帧的内容复杂度为:
Num
∑ / _ slice _ content _ complexity k f _ frame _ content _ complexity =―
Num
具体用于: 根据所述视频帧的内容复杂度确定所述视频的内容复杂度
19、 如权利要求 18所述的设备, 其特征在于,
所述确定单元具体用于: 确定所述视频的内容复杂度为:
Figure imgf000031_0001
f _ video _ content _ complexity =
M
其中, 所述视频流包括 M个视频帧, M为正整数, j为正整数且取值从 1到 M。
20、 如权利要求 19所述的设备, 其特征在于,
所述预测单元具体用于: 预测视频编码质量为: d _ compression _ quality _ value =
a
a α - α5 · f _ video _ content _ complexity
其中, a5和 a6均为常数。
21、 如权利要求 20所述的设备, 其特征在于, 所述 a2取值为: 视频编码质量范围的最大值与最小值之差。
22、 如权利要求 20或 21所述的设备, 其特征在于, 所述 1取值为: 视频编码质量范围的最小值。
23、 如权利要求 18-22任一项所述的设备, 其特征在于, 还包括, 对所 述视频的内容复杂度进行调整:
f _ video _ content _ complexity _ n =
Figure imgf000031_0002
将调整后的 f— video— content— complexity— n作为所述视频的内容复杂度, s和 numl均为常数。
24、 如权利要求 23所述的设备, 其特征在于, 所述 s取值为 1.0, 所述 numl取值为 60.0, 所述 f—vidio— content— complexity— n取值范围为 [0.0, 1 ·0]。
25、 如权利要求 19-24任一项所述的方法, 其特征在于, 所述 Μ个视频 帧均没有丟包。
26、 如权利要求 19-25任一项所述的设备, 其特征在于, 所述视频帧包 括帧内编码帧和 /或帧间编码帧。
27、 如权利要求 18-26任一项所述的设备, 其特征在于, Num为所述视 频帧中条带的总数。
28、 如权利要求 17-27任一项所述的设备, 其特征在于, 所述 N1为正 确解析出的条带的量化参数的数目,所述正确解析出的条带的量化参数表示 在条带的第一个传输包至用于获取所述条带的量化参数的条带量化参数差 量所在的传输包都不丟失的情况下, 解析出的条带的量化参数为正确的。
PCT/CN2013/080882 2013-02-06 2013-08-06 视频编码质量的评估方法及设备 Ceased WO2014121591A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
EP13874341.4A EP2919461B1 (en) 2013-02-06 2013-08-06 Method and device for evaluating video encoding quality
ES13874341.4T ES2622027T3 (es) 2013-02-06 2013-08-06 Método y dispositivo para evaluar calidad de codificación de vídeo
US14/751,373 US9774855B2 (en) 2013-02-06 2015-06-26 Method and device for assessing video encoding quality

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201310048015.0 2013-02-06
CN201310048015.0A CN103747237B (zh) 2013-02-06 2013-02-06 视频编码质量的评估方法及设备

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US14/751,373 Continuation US9774855B2 (en) 2013-02-06 2015-06-26 Method and device for assessing video encoding quality

Publications (1)

Publication Number Publication Date
WO2014121591A1 true WO2014121591A1 (zh) 2014-08-14

Family

ID=50504221

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2013/080882 Ceased WO2014121591A1 (zh) 2013-02-06 2013-08-06 视频编码质量的评估方法及设备

Country Status (5)

Country Link
US (1) US9774855B2 (zh)
EP (1) EP2919461B1 (zh)
CN (1) CN103747237B (zh)
ES (1) ES2622027T3 (zh)
WO (1) WO2014121591A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111602105A (zh) * 2018-01-22 2020-08-28 苹果公司 用于呈现合成现实伴随内容的方法和设备
CN112913238A (zh) * 2018-10-31 2021-06-04 Ati科技无限责任公司 用于低等待时间视频写码的有效量化参数预测方法
CN116137658A (zh) * 2021-11-17 2023-05-19 北京猿力未来科技有限公司 视频编码方法及装置

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105336338B (zh) * 2014-06-24 2017-04-12 华为技术有限公司 音频编码方法和装置
US10163030B2 (en) * 2016-05-20 2018-12-25 Gopro, Inc. On-camera image processing based on image activity data
CN109819263B (zh) * 2017-11-22 2022-08-19 腾讯科技(深圳)有限公司 视频编码方法、装置、计算机设备及存储介质
US10798399B1 (en) * 2017-12-11 2020-10-06 Amazon Technologies, Inc. Adaptive video compression
CN109286812B (zh) * 2018-10-24 2020-06-05 华中科技大学 一种hevc视频质量估计方法
CN109660812A (zh) * 2018-11-12 2019-04-19 北京达佳互联信息技术有限公司 复杂度和码率的确定方法、装置及计算机可读存储介质
CN111385576B (zh) * 2018-12-28 2021-08-10 北京字节跳动网络技术有限公司 视频编码方法、装置、移动终端及存储介质
CN111327950B (zh) * 2020-03-05 2021-08-06 腾讯科技(深圳)有限公司 一种视频转码方法及装置
CN113395518B (zh) * 2021-08-13 2022-03-15 杭州微帧信息科技有限公司 一种视频编码自适应量化死区的方法
CN115225961B (zh) * 2022-04-22 2024-01-16 上海赛连信息科技有限公司 一种无参考网络视频质量评价方法和装置
CN115174897B (zh) * 2022-07-28 2025-02-14 北京达佳互联信息技术有限公司 视频质量预测方法、装置、电子设备、存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040081361A1 (en) * 2002-10-29 2004-04-29 Hongyi Chen Method for performing motion estimation with Walsh-Hadamard transform (WHT)
CN1816141A (zh) * 2005-02-01 2006-08-09 华为技术有限公司 一种视频编码速率控制的方法
US20110243233A1 (en) * 2010-04-05 2011-10-06 Samsung Electronics Co., Ltd. Method and apparatus for encoding and decoding video
US20110292997A1 (en) * 2009-11-06 2011-12-01 Qualcomm Incorporated Control of video encoding based on image capture parameters
CN102740108A (zh) * 2011-04-11 2012-10-17 华为技术有限公司 一种视频数据质量评估方法和装置
CN102868907A (zh) * 2012-09-29 2013-01-09 西北工业大学 部分参考视频质量客观评价方法

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2009260940A (ja) * 2008-03-21 2009-11-05 Nippon Telegr & Teleph Corp <Ntt> 映像品質客観評価方法、映像品質客観評価装置、およびプログラム
EP2106154A1 (en) * 2008-03-28 2009-09-30 Deutsche Telekom AG Audio-visual quality estimation
EP2296379A4 (en) 2008-07-21 2011-07-20 Huawei Tech Co Ltd METHOD, SYSTEM AND DEVICE FOR EVALUATING A VIDEO QUALITY
CN101695141B (zh) * 2009-10-20 2012-05-23 浙江大学 一种视频质量评估方法和装置
CN102685548B (zh) * 2012-05-29 2015-09-30 公安部第三研究所 视频质量的无参评估方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040081361A1 (en) * 2002-10-29 2004-04-29 Hongyi Chen Method for performing motion estimation with Walsh-Hadamard transform (WHT)
CN1816141A (zh) * 2005-02-01 2006-08-09 华为技术有限公司 一种视频编码速率控制的方法
US20110292997A1 (en) * 2009-11-06 2011-12-01 Qualcomm Incorporated Control of video encoding based on image capture parameters
US20110243233A1 (en) * 2010-04-05 2011-10-06 Samsung Electronics Co., Ltd. Method and apparatus for encoding and decoding video
CN102740108A (zh) * 2011-04-11 2012-10-17 华为技术有限公司 一种视频数据质量评估方法和装置
CN102868907A (zh) * 2012-09-29 2013-01-09 西北工业大学 部分参考视频质量客观评价方法

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111602105A (zh) * 2018-01-22 2020-08-28 苹果公司 用于呈现合成现实伴随内容的方法和设备
CN111602105B (zh) * 2018-01-22 2023-09-01 苹果公司 用于呈现合成现实伴随内容的方法和设备
CN112913238A (zh) * 2018-10-31 2021-06-04 Ati科技无限责任公司 用于低等待时间视频写码的有效量化参数预测方法
CN116137658A (zh) * 2021-11-17 2023-05-19 北京猿力未来科技有限公司 视频编码方法及装置

Also Published As

Publication number Publication date
CN103747237B (zh) 2015-04-29
ES2622027T3 (es) 2017-07-05
EP2919461A1 (en) 2015-09-16
EP2919461B1 (en) 2017-01-11
CN103747237A (zh) 2014-04-23
US9774855B2 (en) 2017-09-26
EP2919461A4 (en) 2015-09-16
US20150296208A1 (en) 2015-10-15

Similar Documents

Publication Publication Date Title
WO2014121591A1 (zh) 视频编码质量的评估方法及设备
JP5400876B2 (ja) ビデオ符号化のための、スライス依存性に基づくレート制御モデル適合化
US10645382B2 (en) Video processing method, encoding device, and decoding device
EP2194720A1 (en) Method and apparatus for encoder assisted-frame rate up conversion (EA-FRUC) for video compression
Yim et al. Evaluation of temporal variation of video quality in packet loss networks
CN106210721B (zh) 一种hevc快速码率转码方法
CN104539948A (zh) 视频处理系统和视频处理方法
Adeyemi-Ejeye et al. Packet loss visibility across SD, HD, 3D, and UHD video streams
JP4861371B2 (ja) 映像品質推定装置、方法、およびプログラム
McDonagh et al. Quality-oriented scalable video delivery using H. 264 SVC on an LTE network
WO2012061258A2 (en) Parametric bit rate model for frame-level rate control in video coding
JP5756154B2 (ja) 映像符号化パラメータ算出装置、映像符号化パラメータ算出方法及びプログラム
AU2013307970A1 (en) Method and apparatus for estimating content complexity for video quality assessment
CN103918258A (zh) 减少视频编码中的数据量
US20130077674A1 (en) Method and apparatus for encoding moving picture
Migliorini et al. Performance evaluation of H. 264/SVC video streaming over mobile WiMAX
US9723266B1 (en) Lightweight content aware bit stream video quality monitoring service
Migliorini et al. QoE-oriented performance evaluation of video streaming over WiMAX
Liu et al. Real-time video quality monitoring
Park GOP level rate-control for real-time video transmission
CN106131551B (zh) 一种用于移动视频直播的分级编码方法
Flayyih et al. H. 264 Video Transmission with High Quality and Low Bitrate over Wireless Network
Sankisa et al. Distortion estimation using structural similarity for video transmission over wireless networks
Al-khrayshah et al. A real-time SNR scalable transcoder for MPEG-2 video streams
NS et al. Novel Intra-Prediction Framework for H. 264 Video Compression using Decision and Prediction Mode

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: 13874341

Country of ref document: EP

Kind code of ref document: A1

REEP Request for entry into the european phase

Ref document number: 2013874341

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 2013874341

Country of ref document: EP

NENP Non-entry into the national phase

Ref country code: DE