WO2009116666A1 - 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 - Google Patents
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 Download PDFInfo
- Publication number
- WO2009116666A1 WO2009116666A1 PCT/JP2009/055679 JP2009055679W WO2009116666A1 WO 2009116666 A1 WO2009116666 A1 WO 2009116666A1 JP 2009055679 W JP2009055679 W JP 2009055679W WO 2009116666 A1 WO2009116666 A1 WO 2009116666A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- bit
- slice
- video
- information
- frame
- Prior art date
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N17/00—Diagnosis, testing or measuring for television systems or their details
- H04N17/004—Diagnosis, testing or measuring for television systems or their details for digital television systems
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/174—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a slice, e.g. a line of blocks or a group of blocks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—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
-
- 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
- H04N19/577—Motion compensation with bidirectional frame interpolation, i.e. using B-pictures
Definitions
- the present invention provides a method for objectively deriving subjective quality from encoded bit string information without performing a subjective quality evaluation experiment when estimating the quality (subjective quality) experienced by humans when viewing a video.
- the present invention relates to a video quality objective evaluation method, a video quality objective evaluation apparatus, and a program for video quality degradation caused by computerization.
- the conventional technology aims to build a technology that objectively evaluates the video quality while reducing the amount of calculation with high accuracy.
- the technique described in Document 1 estimates subjective quality assuming an average scene, it cannot take into account fluctuations in subjective quality due to differences in scenes, and achieves subjective quality estimation with good accuracy. There was a problem that it was not possible.
- the present invention receives a bit sequence of a video encoded using motion compensated inter-frame prediction and DCT transform, and inputs information included in the received bit sequence to input a predetermined bit sequence. And a subjective quality estimation step for performing a calculation for estimating the subjective quality of the video based on the calculation result of the calculation step.
- the present invention is a coding method using motion compensation inter-frame prediction and DCT transform, which are currently mainstream, and particularly H.264.
- the video quality can be estimated with a small amount of calculation, and the motion vector and the content of the DCT coefficient, which are parameters that can take into account the difference in scenes existing in the bit stream, By using the data amount for the estimation of the subjective quality, it becomes possible to estimate the subjective quality of the video with high accuracy.
- the calculation step extracts the quantization information included in the bit string, calculates the statistic of the quantization information (for example, the minimum value of the quantization parameter of the H.264 method),
- the quality estimation step performs an operation for estimating the subjective quality of the video based on the statistic of the quantization information (for example, the minimum value of the quantization parameter of the H.264 method).
- the calculation step extracts motion vector information included in the bit string, calculates a motion vector statistic (for example, kurtosis of the magnitude of the vector) from the extracted motion vector information, and a subjective quality estimation step. Estimates the subjective quality of the video based on the statistics of quantization information (for example, the minimum value of the quantization parameter of the H.264 method) and the statistics of motion vectors (for example, the kurtosis of the vector size). Perform the operation.
- a motion vector statistic for example, kurtosis of the magnitude of the vector
- the calculation step extracts information on the I slice, P slice, and B slice included in the bit string, and statistics on the I slice, P slice, and B slice based on the extracted information on the I slice, P slice, and B slice.
- the information is calculated, and the subjective quality estimation step is performed based on the statistical amount of quantization information (for example, the minimum value of the quantization parameter of the H.264 method) and the statistical information of the I slice, P slice, and B slice. Performs operations to estimate quality.
- the calculation step information used for predictive coding, information used for transform coding, and information used for control of coding included in the bit string is extracted, and prediction coding is performed from the extracted information. It is also possible to calculate the bit amount used for the bit rate, the bit amount used for transform coding, and the bit amount used for coding control.
- the subjective quality estimation step indicates the subjective result of the video based on the bit amount used for predictive coding, the bit amount used for transform coding, and the bit amount used for coding control, which shows the calculation result of the calculation step. Performs operations to estimate quality.
- the calculation step extracts motion vector information included in the bit string, calculates a motion vector statistic (for example, kurtosis of the vector magnitude) from the extracted motion vector information, and performs subjective quality estimation.
- the step is based on the amount of bits used for predictive coding, the amount of bits used for transform coding, the amount of bits used for control of coding and the statistics of motion vectors (eg kurtosis of vector magnitude). Performs computation to estimate the subjective quality of the video.
- the calculation step extracts information on the I slice, P slice, and B slice included in the bit string, and statistics on the I slice, P slice, and B slice based on the extracted information on the I slice, P slice, and B slice.
- the subjective quality estimation step calculates the bit amount used for predictive coding, the bit amount used for transform coding, the bit amount used for control of coding, and statistical information of I slice, P slice, and B slice. Based on the above, a calculation for estimating the subjective quality of the video is performed.
- an encoding method using DCT transform and motion compensation particularly H.264.
- DCT transform and motion compensation particularly H.264.
- FIG. 1 is a diagram for explaining a method for deriving the minimum value of the quantization parameter in the selected frame.
- FIG. 2 is a diagram illustrating the position of a motion vector derivation target frame.
- FIG. 3 is a diagram illustrating a positional relationship between a motion vector derivation target frame and a reference frame.
- FIG. 4 is a diagram illustrating a positional relationship between a motion vector derivation target frame and a reference frame.
- FIG. 5 is a block diagram showing the configuration of the video quality objective evaluation apparatus.
- FIG. 6 is a functional block diagram showing the configuration of the first embodiment of the present invention.
- FIG. 7 is a functional block diagram showing the configuration of the second embodiment of the present invention.
- FIG. 6 is a functional block diagram showing the configuration of the first embodiment of the present invention.
- FIG. 8 is a functional block diagram showing the configuration of the third embodiment of the present invention.
- FIG. 9 is a functional block diagram showing the configuration of the fourth embodiment of the present invention.
- FIG. 10 is a functional block diagram showing the configuration of the fifth embodiment of the present invention.
- FIG. 11 is a functional block diagram showing the configuration of the sixth embodiment of the present invention.
- FIG. 12 is a functional block diagram showing the configuration of the seventh embodiment of the present invention.
- FIG. 13 is a functional block diagram showing the configuration of the eighth embodiment of the present invention.
- FIG. 14 is a functional block diagram showing the configuration of the ninth embodiment of the present invention.
- FIG. 15 is a flowchart showing the processing operation of the first embodiment of the present invention.
- FIG. 16 is a flowchart showing the processing operation of the second embodiment of the present invention.
- FIG. 15 is a flowchart showing the processing operation of the first embodiment of the present invention.
- FIG. 16 is a flowchart showing the processing operation of the second embodiment of the present invention
- FIG. 17 is a flowchart showing the processing operation of the third embodiment of the present invention.
- FIG. 18 is a flowchart showing the processing operation of the fourth embodiment of the present invention.
- FIG. 19 is a flowchart showing the processing operation of the fifth embodiment of the present invention.
- FIG. 20 is a flowchart showing the processing operation of the sixth embodiment of the present invention.
- FIG. 21 is a flowchart showing the processing operation of the seventh embodiment of the present invention.
- FIG. 22 is a flowchart showing the processing operation of the eighth embodiment of the present invention.
- FIG. 23 is a flowchart showing the processing operation of the ninth embodiment of the present invention.
- FIG. 24A is a diagram illustrating characteristics of quantization coefficients / parameters.
- FIG. 24B is a diagram illustrating characteristics of quantization coefficients / parameters.
- FIG. 25A is a diagram showing a comparison result of accuracy with a general quality estimation model.
- FIG. 25B is a diagram showing a comparison result of accuracy with
- FIG. 5 is a block diagram showing the configuration of the video quality objective evaluation apparatus in the embodiment of the present invention.
- the video quality objective evaluation apparatus 1 includes a reception unit 2, a calculation unit 3, a storage medium 4, and an output unit 5.
- the H.264 encoding device 6 converts the input video into the H.264 format described later. It is encoded by the H.264 system.
- the encoded video bit string is communicated as a transmission packet in the transmission network and transmitted to the video quality objective evaluation apparatus 1.
- the video quality objective evaluation apparatus 1 receives the transmission packet, that is, the encoded bit string at the receiving unit 2. And each function of the calculating part 3 is implement
- the subjective quality of the video is estimated by outputting to.
- a quantization parameter statistic calculation unit 11 and an integration unit 20 are provided.
- the subjective quality EV is estimated using quantization parameter information that is quantization information existing in the bit stream of the evaluation video V encoded by the H.264 method. This method can be applied in principle to a coding method using DCT coefficients and motion compensation.
- the simple flow of this embodiment is as follows. First, in FIG. 6, the encoded bit stream is first input to the quantization parameter statistic calculator 11.
- the quantization parameter statistic calculation unit 11 extracts the quantization parameter from the bit stream and derives the representative value QPmin of the quantization parameter according to the algorithm shown below.
- the integration unit 20 estimates the subjective quality EV of the evaluation video V from the representative value QPmin of the quantization parameter according to the following algorithm.
- the quantization parameter values of all macroblocks (all m) existing in the frame are used to represent quantization parameters for each frame.
- the value QPmin (i) is derived.
- QPmin (i) is derived from the following equation.
- QPij represents the quantization parameter of macroblock number j in frame number i (FIG. 1), and operator
- QP min is derived from the following equation.
- the subjective quality EV of the evaluation video V is estimated using QP min derived as described above.
- Subjective quality EV is derived by the following equation. H. In 264 the system due to the presence of non-linearity between the representative value QP min and subjective quality EV of the quantization parameter, and is obtained by considering the characteristics.
- the coefficients a, b, c, and d are coefficients that are optimized by regression analysis by conducting a subjective evaluation experiment in advance.
- the ACR method shown in Reference 2 ITU-T P.910, “TELEPHONE TRANSMISSION QUALITY, TELEPHONE INSTALLATIONS, LOCAL LINE NETWORKS,” Sep. 1999.
- Reference 3 ITU-T R BT.500, “Methodology for the subjective assessment of the quality of television pictures,” 2002.) and the DSIS method or DSCQS method.
- QP min (i) which is a quantization parameter for each frame
- QP min (i) statistics such as QP ave which is an average value of QP min (i) and QP max which is a maximum value are calculated. It may be used instead of QP min .
- a quantization parameter statistic calculation unit 11 a motion vector statistic calculation unit 12, and an integration unit 20 are provided.
- Subjective quality EV is objectively estimated using motion vector information in addition to quantization parameters used in H.264 encoding. This scheme is applicable in principle to a coding scheme that uses DCT coefficients and motion compensation.
- the encoded bit stream is first input to the quantization parameter statistic calculation unit 11 and the motion vector statistic calculation unit 12.
- the quantization parameter statistic calculation unit 11 extracts the quantization parameter from the bit stream and derives the representative value QPmin of the quantization parameter according to the algorithm shown below.
- the motion vector statistic calculation unit 12 extracts a motion vector from the bitstream and derives a representative value MVkurt of the motion vector according to the algorithm shown below.
- the integration unit 20 estimates the subjective quality EV of the evaluation video V from the representative value QPmin of the quantization parameter and the representative value MVkurt of the motion vector according to the following algorithm.
- QP min shown in the first embodiment is used for EV derivation.
- the representative value of the motion vector will be described with reference to FIG. 2, FIG. 3, and FIG.
- any two reference frames used for derivation of motion vectors can be selected in units of macroblocks and sub-macroblocks, not limited to the front and rear. Therefore, each macroblock / sub-macroblock is projected onto one frame before and after the motion vector derivation target frame so as to normalize the size of the motion vector set for each macroblock / sub-macroblock. Specific processing will be described with reference to FIGS.
- FIG. 3 shows a case where the reference frame of the j-th block MB ij in the motion vector derivation target frame i is behind (p + 1) frames of the frame i.
- a motion vector MV ij exists on a reference frame from a motion vector derivation target frame i, and MV ij is converted to a vector MV from the motion vector derivation target frame i one frame backward as follows. ' Project to ij .
- FIG. 4 shows a case where the reference frame of the j-th block MB ij in the motion vector derivation target frame i is ahead of the (q + 1) frame of the frame i.
- the motion vector MV ij is present on the reference frame from the motion vector derivation target frame i, and the MV ij is moved forward from the motion vector derivation target frame i by one frame MV as follows. ' Project to ij .
- the motion vector set for every macroblock / sub-macroblock j (1 ⁇ j ⁇ x) of the motion vector derivation target frame i can be projected onto a vector on i ⁇ 1 frame. It becomes possible.
- x is the number of macroblocks in frame i.
- the kurtosis Kurt (i) is derived as a statistic of the motion vector derivation target frame i by the following equation.
- various statistics such as an average, maximum value, minimum value, and variance can be used as an alternative. In the following formula
- MV kurt is derived using the representative value MV kurt (i) of the motion vector for each frame derived above.
- MV kurt is derived by the following equation.
- the kurtosis of the motion vector is used to express the distribution of the motion vector, and to quantify the uniform motion and the motion of a specific object in the video.
- a feature quantity (dispersion, skewness, etc.) having a physical meaning similar to this may be used.
- the subjective quality EV of the evaluation video V is estimated using the MV kurt and QP min derived above. EV is derived from the following equation. MV kurt in the following equation represents the size of the vector.
- the coefficients a, b, c, d, e, f, g, h, i, j, k, l, and m are coefficients that are optimized by regression analysis by conducting a subjective evaluation experiment in advance.
- Examples of EV scales include the ACR method shown in Reference 2 and the DSIS method or DSCQS method shown in Reference 3.
- a quantization parameter statistic calculation unit 11 For the evaluation video V encoded in the H.264 format, The subjective quality EV is objectively estimated using the statistical information of the I slice, P slice, and B slice in addition to the quantization parameter used in the H.264 encoding.
- the switching I slice is regarded as an I slice, and the switching P slice is regarded as a P slice. This scheme is applicable in principle to a coding scheme that uses DCT coefficients and motion compensation.
- the encoded bit stream is first input to the quantization parameter statistic calculator 11 and the frame type statistic calculator 13.
- the quantization parameter statistic calculation unit 11 extracts the quantization parameter from the bit stream and derives the representative value QPmin of the quantization parameter according to the algorithm shown below.
- the frame type statistic calculation unit 13 extracts the frame type from the bit stream and derives the frame type statistic R according to the algorithm shown below.
- the integration unit 20 estimates the subjective quality EV of the evaluation video V from the representative value QPmin of the quantization parameter and the frame type statistic R according to the following algorithm.
- QPmin shown in the first embodiment (however, the statistic of the quantization parameter shown in the first embodiment may be used) is used for EV derivation.
- I, P, and B attributes set for each slice S I counted the number of I slices present in the evaluation video, S P counted the number of P slices, and counted the number of B slices.
- S B is derived, and the ratio R SI , R SP , R SB , R SPB of the number of each slice to the number of all slices is derived by the following equation.
- the subjective quality EV of the evaluation video V is estimated using R and QP min derived as described above. EV is derived from the following equation.
- the coefficients a, b, c, d, e, f, g, h, i, j, k, l, and m are coefficients that are optimized by regression analysis by conducting a subjective evaluation experiment in advance.
- Examples of EV scales include the ACR method shown in Reference 2 and the DSIS method or DSCQS method shown in Reference 3.
- a quantization parameter statistic calculation unit 11 a motion vector statistic calculation unit 12, a frame type statistic calculation unit 13, and an integration unit 20 are provided.
- the subjective quality EV is objectively estimated using the motion vector and the information of the I slice, P slice, and B slice in addition to the quantization parameter used in the H.264 encoding.
- the switching I slice is regarded as an I slice
- the switching P slice is regarded as a P slice.
- the encoded bit stream is first input to the quantization parameter statistic calculator 11, the motion vector statistic calculator 12, and the frame type statistic calculator 13.
- the quantization parameter statistic calculation unit 11 extracts the quantization parameter from the bit stream and derives the representative value QPmin of the quantization parameter according to the algorithm shown below.
- the motion vector statistic calculation unit 12 extracts a motion vector from the bit stream and derives a representative value MV kurt of the motion vector according to the algorithm described below.
- the frame type statistic calculation unit 13 extracts the frame type from the bit stream and derives the frame type statistic R according to the algorithm shown below.
- the integrating unit 20 estimates the subjective quality EV of the evaluation video V from the representative value QPmin of the quantization parameter, the representative value MV kurt of the motion vector, and the frame type statistic R according to the following algorithm.
- QP min shown in the first embodiment (however, the statistic of the quantization parameter shown in the first embodiment may be used) is used for EV derivation.
- the motion vector the MV kurt shown in the second embodiment is used to derive the EV.
- the I slice, P slice, and B slice, R shown in the third embodiment is used.
- the subjective quality EV of the evaluation video V is estimated using the derived MV kurt , R, and QP min .
- EV is derived from the following equation.
- MV kurt in the following equation represents the size of the vector.
- coefficients a, b, c, d, e, f, g, h, i, j, k, l, m, o, p, q, r, s, t, u, v, w, x, y Is a coefficient which is optimized by regression analysis by conducting a subjective evaluation experiment in advance.
- Examples of EV scales include the ACR method shown in Reference 2 and the DSIS method or DSCQS method shown in Reference 3.
- a statistic calculation unit 14 and an integration unit 20 of the sum of bit amounts are provided.
- the subjective quality EV is estimated using the bit amount used for predictive coding of the bit stream of the evaluation video V encoded by the H.264 method, the bit amount used for transform coding, and the bit amount used for coding control. .
- This scheme is applicable in principle to a coding scheme that uses DCT coefficients and motion compensation.
- the encoded bit stream is first input to the bit quantity sum statistic calculation unit 14.
- the bit amount sum statistic calculation unit 14 extracts the sum of the bit amounts from the bit stream and derives a representative value Bit max of the sum of the bit amounts according to the following algorithm.
- the integration unit 20 estimates the subjective quality EV of the evaluation video V from the representative value Bit max of the sum of the bit amounts according to the following algorithm.
- Bit max (i) of the sum of bit quanta for each frame is derived using the sum of the amount and the bit amount used for encoding control.
- Bit max (i) is derived from the following equation.
- Bit ij represents the sum of the bit amounts of the macroblock number j in the frame number i, and the operator
- Bit max (i) of the sum of the bit amount for each frame derived above.
- the subjective quality EV of the evaluation video V is estimated using Bit max derived above.
- Subjective quality EV is derived by the following equation. H. In H.264, since there is nonlinearity between the representative value Bit max of the sum of the bit amounts and the subjective quality EV, the characteristics are taken into consideration.
- the coefficients a, b, c, and d are coefficients that are optimized by regression analysis by conducting a subjective evaluation experiment in advance.
- Examples of EV scales include the ACR method shown in the above-mentioned Reference 2 and the DSIS method or DSCQS method shown in the above-mentioned Reference 3.
- Bit max (i) are representative of the sum of the bit amount for each frame, the average value of Bit max (i) Bitave and, Bit min such a minimum value, Bit max of (i) Statistics may be used instead of Bit max .
- a statistic calculation unit 14 for the sum of bit amounts a motion vector statistic calculation unit 12, and an integration unit 20 are provided.
- the subjective quality EV is objectively estimated using the bit amount used for predictive coding of the bitstream used in H.264 coding, the bit amount used for transform coding, and the bit amount used for coding control. To do.
- This scheme is applicable in principle to a coding scheme that uses DCT coefficients and motion compensation.
- the encoded bit stream is first input to the sum statistic calculator 14 and the motion vector statistic calculator 12.
- the bit amount sum statistic calculation unit 14 extracts the sum of the bit amounts from the bit stream and derives a representative value Bit max of the sum of the bit amounts according to the following algorithm.
- the motion vector statistic calculation unit 12 extracts a motion vector from the bit stream and derives a representative value MV kurt of the motion vector according to the algorithm described below.
- the integration unit 20 estimates the subjective quality EV of the evaluation video V from the representative value Bit max of the sum of the bit amounts and the representative value MV kurt of the motion vector according to the following algorithm.
- Bitmax shown in Embodiment 5 is used for EV derivation.
- the subjective quality EV of the evaluation video V is estimated using the representative value MV kurt of all the motion vectors of the evaluation video derived above and the representative value Bit max of the sum of all the bit amounts of the evaluation video. EV is derived from the following equation. MV kurt in the following equation represents the size of the vector.
- the coefficients a, b, c, d, e, f, g, h, i, j, k, l, and m are coefficients that are optimized by regression analysis by conducting a subjective evaluation experiment in advance.
- Examples of EV scales include the ACR method shown in Reference 2 and the DSIS method or DSCQS method shown in Reference 3.
- a statistic calculation unit 14 for the sum of bit amounts, a frame type statistic calculation unit 13 and an integration unit 20 are provided.
- I slice, P slice, B Subjective quality EV is objectively estimated using slice statistical information.
- the switching I slice is regarded as an I slice, and the switching P slice is regarded as a P slice. This method is applicable in principle to a coding method using DCT coefficients and motion compensation.
- the encoded bit stream is first input to the sum statistic calculator 14 and the frame type statistic calculator 13.
- the bit amount sum statistic calculation unit 14 extracts the sum of the bit amounts from the bit stream and derives a representative value Bit max of the sum of the bit amounts according to the following algorithm.
- the frame type statistic calculation unit 13 extracts the frame type from the bit stream and derives the frame type statistic R according to the algorithm shown below.
- the integration unit 20 estimates the subjective quality EV of the evaluation video V from the representative value Bit max of the sum of the bit amounts and the frame type statistic R according to the following algorithm.
- Bit max shown in Embodiment 5 is used for EV derivation. .
- the number of S I and P slices that count the number of I slices existing in the evaluation video is counted.
- S B obtained by counting the number of S P and B slices is derived, and ratios R SI , R SP , R SB , and R SPB of the number of each slice to the number of all slices are derived as parameters. Then, using these parameters, the correlation with the subjective quality previously derived in the subjective quality evaluation experiment by the regression analysis is compared, and the parameter with the highest subjective quality estimation accuracy is defined as R.
- the subjective quality EV of the evaluation video V is estimated using the parameter R and Bit max derived above. EV is derived from the following equation.
- the coefficients a, b, c, d, e, f, g, h, i, j, k, l, and m are coefficients that are optimized by regression analysis by conducting a subjective evaluation experiment in advance.
- Examples of EV scales include the ACR method shown in Reference 2 and the DSIS method or DSCQS method shown in Reference 3.
- a statistic calculation unit 14 for the sum of bit amounts a motion vector statistic calculation unit 12, a frame type statistic calculation unit 13, and an integration unit 20 are provided.
- P Subjective quality EV is objectively estimated using information on slices and B slices.
- the switching I slice is regarded as an I slice, and the switching P slice is regarded as a P slice. This scheme is applicable in principle to a coding scheme that uses DCT coefficients and motion compensation.
- the encoded bitstream is first input to the bit amount sum statistic calculation unit 14, the motion vector statistic calculation unit 12, and the frame type statistic calculation unit 13.
- the bit amount sum statistic calculation unit 14 extracts the sum of the bit amounts from the bit stream and derives a representative value Bit max of the sum of the bit amounts according to the following algorithm.
- the motion vector statistic calculation unit 12 extracts a motion vector from the bit stream and derives a representative value MV kurt of the motion vector according to the algorithm described below.
- the frame type statistic calculation unit 13 extracts the frame type from the bit stream and derives the frame type statistic R according to the algorithm shown below.
- the integrating unit 20 estimates the subjective quality EV of the evaluation video V from the representative value Bit max of the sum of the bit amounts, the representative value MV kurt of the motion vector, and the frame type statistic R according to the following algorithm.
- Bit max shown in Embodiment 5 is used for EV derivation.
- the motion vector the MV kurt shown in the second embodiment is used to derive the EV.
- the I slice, P slice, and B slice, R shown in the third embodiment is used.
- the subjective quality EV of the evaluation video V is estimated using the derived MV kurt , R, and QP min .
- EV is derived from the following equation.
- MV kurt in the following equation represents the size of the vector.
- coefficients a, b, c, d, e, f, g, h, i, j, k, l, m, o, p, q, r, s, t, u, v, w, x, y Is a coefficient which is optimized by regression analysis by conducting a subjective evaluation experiment in advance.
- Examples of EV scales include the ACR method shown in Reference 2 and the DSIS method or DSCQS method shown in Reference 3.
- a statistic calculation unit 15 for the sum of bit amounts of I slices, P slices, and B slices, and a statistic calculation unit for quantization information of I slices, P slices, and B slices 16 and a subjective quality estimation unit 17 are provided.
- the amount of bits used for predictive encoding of I slice, P slice, and B slice, the amount of bits used for transform encoding, and used for encoding control is objectively estimated using the quantization parameters (quantization information) of the I slice, P slice, and B slice.
- the switching I slice is regarded as an I slice
- the switching P slice is regarded as a P slice.
- the encoded bit stream is first input to the bit amount sum calculation unit 15 of I slice, P slice, and B slice.
- the calculation unit 15 derives bit amounts of I slices, P slices, and B slices by dividing them into motion vectors, quantization coefficients, and encoding control information.
- the I-slice, P-slice, and B-slice quantization information statistic calculation unit 16 extracts the I-slice, P-slice, and B-slice quantization information, and follows the algorithm shown below to obtain I-slice / P-slice. Deriving statistics QP min (I), QP min (P), and QP min (B) of quantization information of B slices
- the subjective quality estimation unit 17 estimates the subjective quality EV of the evaluation video V according to the following algorithm using the statistics QP min (I), QP min (P), QP min (B), and the like. These flows are shown in the flowchart of FIG. H.
- the specification of the H.264 bit string is described in Reference 1, and The bit amount used for predictive coding of a bitstream set for each macroblock according to the H.264 specification, the bit amount used for transform coding, the bit amount used for coding control, and the I set for each slice -Extract P / B attribute information.
- the bit amount used for predictive coding of each of the I slice, P slice, and B slice, the bit amount used for transform coding, and the bit amount used for control of coding are used for predictive coding of the I slice.
- the amount of bits, the amount of bits used for transform coding, and the amount of bits used for coding control are Bit pred (I), Bit res (I), and Bit other (I), respectively, for predictive coding of P slices.
- the amount of bits used, the amount of bits used for transform encoding, and the amount of bits used for control of encoding are Bit pred (P), Bit res (P), and Bit other (P), respectively.
- bit amount used for encoding is respectively Bit pred (B), Bit res (B), and Bit other (B).
- each bit amount may be the bit amount of all the corresponding slices of the evaluation video, or may be the bit amount of a slice existing within a specific time.
- the values Bit pred (BP), Bit res (BP), and Bit other (BP) are defined, and are derived by the following expressions.
- Bitp red (BP) Bit pred (B) + Bit pred (P)
- Bit res (BP) Bit res (B) + Bit res (P)
- Bit other (BP) Bit other (B) + Bit other (P)
- the QP min derivation process for each slice shown in the first embodiment is applied to only the QP min (I) ⁇ P slice only, and only the QP min (P) ⁇ B slice is applied. QP min (B) applied to is used.
- the I, P, and B attributes are determined for all slices constituting the evaluation video V, and the quantization information values of all macroblocks (all m) existing in the slice for each slice.
- QPij represents quantization information of macroblock number j in slice number i (FIG. 1).
- a value of QP min (BP) is defined and derived by the following formula.
- QP min (BP) (QP min (B) + QP min (P)) / 2
- Subjective quality EV is derived by the following equation. H. In the H.264 system, since nonlinearity exists between the bit amount of the I / P / B slice, the representative value of the quantization information, and the subjective quality EV, the characteristics are taken into consideration.
- the coefficients a, b, c, d, e, f, g, h, i, j, k, l, m, n, and o are coefficients that are optimized by regression analysis by conducting a subjective evaluation experiment in advance. is there.
- Examples of EV scales include the ACR method shown in Reference 2 and the DSIS method or DSCQS method shown in Reference 3.
- Bit other (I) and Bit other (BP) may be used, and subjectivity may be derived by combining each case and applying various statistical operations such as summation, average, and variance, and performing superposition.
- Bit other (I) and Bit other (BP) may be used, and subjectivity may be derived by combining each case and applying various statistical operations such as summation, average, and variance, and performing superposition.
- the above non-linearity may be considered by using a logarithmic function, a polynomial function, or their reciprocal number instead of the exponential function.
- the calculation is performed in units of slices, but it is possible to change the calculation unit in units of macroblocks, frames, GoPs, entire images, and the like.
- FIGS. 24A and 24B show a situation in which the subjective quality is saturated in a region where QP min is small, the subjective quality is rapidly changed in a region where QP min is intermediate, and the subjective quality is saturated in a region where QP min is large.
- FIG. 24B represents the relationship between the bit rate and the subjective quality of scenes 1, 2, 3, and 4 in order from the top, and indicates that the subjective quality varies depending on the encoding difficulty.
- FIG. 25A the estimation results when the subjective quality is estimated by applying the average and standard deviation, which are general statistics, and when the subjective quality is estimated by the model shown in the present invention are shown in FIG. 25A. , Shown in FIG. 25B.
- the horizontal axis represents the subjective quality acquired in the subjective quality evaluation experiment
- the vertical axis represents the objective quality obtained by estimating the subjective quality.
- the general model using the average and the standard deviation does not take into account the saturation characteristics shown in FIG. 24 and causes a deterioration in estimation accuracy.
- the present invention can accurately consider this.
- the estimation accuracy can be improved.
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)
- Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)
Abstract
Description
図5は、本発明の実施の形態における映像品質客観評価装置の構成を示すブロック図である。本映像品質客観評価装置1は、図5に示すように、受信部2、演算部3、記憶媒体4および出力部5からなる。ここで、図5に示すH.264符号化装置6は、入力映像を後述するH.264方式により符号化する。そして、符号化された映像ビット列は伝送網内を伝送パケットとして通信され映像品質客観評価装置1に送信される。
実施の形態1では、図6に示すように、量子化パラメータ統計量計算部11と統合部20とが設けられ、H.264方式で符号化された評価映像Vのビットストリーム内に存在する量子化情報である量子化パラメータの情報を使って主観品質EVを推定する。本方式は、DCT係数と動き補償を使う符号化方式には原理的に適用可能である。
上記処理によりフレーム番号i中で最も値が小さい量子化パラメータの導出がなされる。これは量子化パラメータがより小さくなると該当マクロブロックに対してより細かい量子化が適用されることから、上記処理は最も細かい量子化を行うマクロブロックを導出することになる。絵柄が複雑な映像ほどきめ細かい量子化が必要になるため、上記処理はフレーム番号i中で最も複雑な絵柄となるマクロブロックの特定を目的としている。
以上で導出したQPminを用いて、評価映像Vの主観品質EVを推定する。主観品質EVは以下の式で導出される。H.264では方式では量子化パラメータの代表値QPminと主観品質EVの間に非線形性が存在するため、その特性を考慮したものとなっている。
実施の形態2では、図7に示すように、量子化パラメータ統計量計算部11、動きベクトル統計量計算部12および統合部20が設けられ、H.264方式で符号化された評価映像Vに対して、H.264方式の符号化で利用される量子化パラメータに加えて動きベクトルの情報を使って主観品質EVを客観的に推定する。本方式は、原理的にはDCT係数と動き補償を使う符号化方式に適用可能である。
一方、動きベクトルの代表値については、図2、図3、図4を用いて説明する。図2に示すように、H.264方式では動きベクトルの導出に用いる参照フレームを、前方、後方に限らず、マクロブロック・サブマクロブロック単位で任意の2つ選ぶことが可能である。そこで各マクロブロック・サブマクロブロック毎に設定される動きベクトルの大きさを正規化するように、動きベクトルの導出対象フレームの前方・後方1フレームに各マクロブロック・サブマクロブロックを射影する。具体的な処理を図3、図4で説明する。
下式中の
ここで動きベクトルの尖度を用いたのは、動きベクトルの分布を表現するためであり、映像で一様な動きや特定の物体の動きを定量化するためである。これと類似の物理的意味を持った特徴量(分散や歪度等)を使っても良い。
実施の形態3では、図8に示すように、量子化パラメータ統計量計算部11、フレーム種別統計量計算部13および統合部20が設けられ、H.264方式で符号化された評価映像Vに対して、H.264方式の符号化で利用される量子化パラメータに加えてIスライス、Pスライス、Bスライスの統計情報を使って主観品質EVを客観的に推定する。なお、スイッチングIスライスはIスライス、スイッチングPスライスはPスライスとみなす。本方式は、原理的にはDCT係数と動き補償を使う符号化方式に適用可能である。
一方、スライス毎に設定されるI・P・B属性については、評価映像中に存在するIスライスの数をカウントしたSI、Pスライスの数をカウントしたSP、Bスライスの数をカウントしたSBを導出し、以下の式で全てのスライスの数に対する各スライスの数の割合RSI、RSP、RSB、RSPBを導出する。これは基本的にPスライスやBスライスなどの他スライスからの差分となるスライスを増やすと原理的に1枚当たりのスライス品質が向上し、逆にIスライスを増やすと1枚当たりのスライス品質が低下するため、全体のスライス数に対する各スライスの割合が品質に密接に関係するために、このようなパラメータを導入した。ただし、スライスの代わりに、I・P・B属性のフレームまたはブロックを用いて上記処理を実行しても良い。
実施の形態4では、図9に示すように、量子化パラメータ統計量計算部11、動きベクトル統計量計算部12、フレーム種別統計量計算部13および統合部20が設けられ、H.264方式で符号化された評価映像Vに対して、H.264方式の符号化で利用される量子化パラメータに加えて動きベクトルとIスライス、Pスライス、Bスライスの情報を使って主観品質EVを客観的に推定する。なお、スイッチングIスライスはIスライス、スイッチングPスライスはPスライスとみなす。
動きベクトルについては、実施の形態2で示したMVkurtをEVの導出に利用する。
Iスライス、Pスライス、Bスライスについては実施の形態3で示したRを利用する。
実施の形態5では、図10に示すように、ビット量の和の統計量計算部14と統合部20が設けられ、H.264方式で符号化された評価映像Vのビットストリームの予測符号化に使われるビット量、変換符号化に使われるビット量、符号化の制御に使われるビット量を用いて主観品質EVを推定する。本方式は、原理的にはDCT係数と動き補償を使う符号化方式に適用可能である。
上記処理によりフレーム番号i中で最も値が大きいビット量の和の導出がなされる。これはビット量の和がより大きくなると該当マクロブロックに対してよりビット量を割り振る符号化処理が適用されることから、上記処理は効率的な処理が難しいマクロブロックのビット量の和を導出することになる。
以上で導出したBitmaxを用いて、評価映像Vの主観品質EVを推定する。主観品質EVは以下の式で導出される。H.264では方式ではビット量の和の代表値Bitmaxと主観品質EVの間に非線形性が存在するため、その特性を考慮したものとなっている。
実施の形態6では、図11に示すように、ビット量の和の統計量計算部14、動きベクトル統計量計算部12および統合部20が設けられ、H.264方式で符号化された評価映像Vに対して、H.264方式の符号化で利用されるビットストリームの予測符号化に使われるビット量、変換符号化に使われるビット量、符号化の制御に使われるビット量を用いて主観品質EVを客観的に推定する。本方式は、原理的にはDCT係数と動き補償を使う符号化方式に適用可能である。
一方、動きベクトルの代表値の導出については、既に実施の形態2で図2、図3、図4を用いて説明したとおりであるので省略する。
実施の形態7では、図12に示すように、ビット量の和の統計量計算部14、フレーム種別統計量計算部13および統合部20が設けられ、H.264方式で符号化された評価映像Vに対して、H.264方式の符号化で利用されるビットストリームの予測符号化に使われるビット量、変換符号化に使われるビット量、および符号化の制御に使われるビット量に加えてIスライス、Pスライス、Bスライスの統計情報を用いて主観品質EVを客観的に推定する。なお、スイッチングIスライスはIスライス、スイッチングPスライスはPスライスとみなす。本方式は、原理的にはDCT係数と動き補償を使う符号化方式には適用可能である。
実施の形態8では、図13に示すように、ビット量の和の統計量計算部14、動きベクトル統計量計算部12、フレーム種別統計量計算部13および統合部20が設けられ、H.264方式で符号化された評価映像Vに対して、H.264方式の符号化で利用されるビットストリームの予測符号化に使われるビット量、変換符号化に使われるビット量、および符号化の制御に使われるビット量に加えて動きベクトルとIスライス、Pスライス、Bスライスの情報を使って主観品質EVを客観的に推定する。なお、スイッチングIスライスはIスライス、スイッチングPスライスはPスライスとみなす。本方式は、原理的にはDCT係数と動き補償を使う符号化方式に適用可能である。
動きベクトルについては、実施の形態2で示したMVkurtをEVの導出に利用する。
Iスライス、Pスライス、Bスライスについては実施の形態3で示したRを利用する。
実施の形態9では、図14に示すように、Iスライス・Pスライス・Bスライスのビット量の和の統計量計算部15、Iスライス・Pスライス・Bスライスの量子化情報の統計量計算部16および主観品質推定部17が設けられ、H.264方式で符号化された評価映像Vに対して、H.264方式の符号化で利用されるビットストリームのうち、Iスライス・Pスライス・Bスライスの各々の予測符号化に使われるビット量、変換符号化に使われるビット量、符号化の制御に使われるビット量に加えて、Iスライス・Pスライス・Bスライスの量子化パラメータ(量子化情報)を使って主観品質EVを客観的に推定するものである。なお、スイッチングIスライスはIスライス、スイッチングPスライスはPスライスとみなす。
Bitres(BP)=Bitres(B)+Bitres(P)
Bitother(BP)=Bitother(B)+Bitother(P)
量子化情報については、実施の形態1で示したスライス毎のQPminの導出プロセスを、Iスライスのみに適用したQPmin(I)・Pスライスのみに適用したQPmin(P)・Bスライスのみに適用したQPmin(B)を利用する。
上記処理によりスライス番号i中で最も値が小さい量子化情報の導出がなされる。これは量子化情報がより小さくなると該当マクロブロックに対してより細かい量子化が適用されることから、上記処理は最も細かい量子化を行うマクロブロックを導出することになる。絵柄が複雑な映像ほどきめ細かい量子化が必要になるため、上記処理はスライス番号i中で最も複雑な絵柄となるマクロブロックの特定を目的としている。
なお、QPmin(i)の代わりにQPave(i)等の平均や最小や最大等の平均や最小や最大等の別のパラメータを以下の処理で利用することができる。QPave(i)は以下の式で導出される。
以上で導出したスライス毎の量子化情報の代表値QPmin(i)を用いて、次に評価映像の全ての量子化情報の代表値QPminを導出する。QPminは以下の式で導出される。
QPmin(BP)=(QPmin(B)+QPmin(P))/2
次に、評価映像Vの主観品質EVを推定する。主観品質EVは以下の式で導出される。H.264方式ではI・P・Bスライスのビット量と量子化情報の代表値と主観品質EVの間に非線形性が存在するため、その特性を考慮したものとなっている。
また、Bitres(I)・Bitres(BP)の代わりに、Bitpred(I)・Bitpred(BP)や以下で定義するビット量の割合Rres(I)・Rres(BP)、またBitother(I)・Bitother(BP)を用いることもでき、各場合を組み合わせて和算・平均・分散等の各種統計演算を適用して重ね合わせを行って主観品質を導出しても良い。以上の式について、指数関数の代わりに対数関数や多項式関数や、それらの逆数により上記非線形性を考慮しても良い。
Claims (15)
- 映像の主観品質を評価する評価方法であって、
動き補償フレーム間予測とDCT変換もしくはウェーブレット変換等のその他の直交変換を用いて符号化された前記映像のビット列を受信する受信ステップと、
受信したビット列に含まれる情報を入力して所定の演算を行う演算ステップと、
前記演算ステップの演算結果に基づいて前記映像の主観品質を推定する演算を行う主観品質推定ステップと
を有することを特徴とする映像品質客観評価方法。 - 請求項1において、
前記演算ステップは、前記ビット列に含まれる量子化情報を抽出して、前記量子化情報の統計量を演算し、
前記主観品質推定ステップは、前記量子化情報の統計量に基づいて前記映像の主観品質を推定する演算を行うことを特徴とする映像品質客観評価方法。 - 請求項2において、
前記演算ステップは、前記ビット列に含まれる動きベクトルの情報を抽出して、抽出した動きベクトルの情報から動きベクトルの統計量を演算し、
前記主観品質推定ステップは、前記量子化情報の統計量と前記動きベクトルの統計量に基づいて前記映像の主観品質を推定する演算を行うことを特徴とする映像品質客観評価方法。 - 請求項2において、
前記演算ステップは、前記ビット列に含まれる動き補償フレーム間予測であるI(イントラ)フレームまたはスライスまたはブロック、P(前方向予測)フレームまたはスライスまたはブロックおよびB(双方向、双予測)フレームまたはスライスまたはブロックの情報を抽出するとともに、抽出したI(イントラ)フレームまたはスライスまたはブロック、P(前方向予測)フレームまたはスライスまたはブロックおよびB(双方向、双予測)フレームまたはスライスまたはブロックの情報に基づいてI(イントラ)フレームまたはスライスまたはブロック、P(前方向予測)フレームまたはスライスまたはブロックおよびB(双方向、双予測)フレームまたはスライスまたはブロックの統計情報を演算し、
前記主観品質推定ステップは、前記量子化情報の統計量と、前記I(イントラ)フレームまたはスライスまたはブロック、P(前方向予測)フレームまたはスライスまたはブロックおよびB(双方向、双予測)フレームまたはスライスまたはブロックの各統計情報に基づいて前記映像の主観品質を推定する演算を行うことを特徴とする映像品質客観評価方法。 - 請求項1において、
前記演算ステップは、前記ビット列に含まれる、予測符号化に用いられる情報、変換符号化に用いられる情報および符号化の制御に用いられる情報を抽出して、抽出したこれらの情報から、予測符号化に用いられるビット量、変換符号化に用いられるビット量および符号化の制御に用いられるビット量を演算し、
前記主観品質推定ステップは、前記演算ステップの演算結果を示す、予測符号化に用いられるビット量、変換符号化に用いられるビット量および符号化の制御に用いられるビット量に基づいて前記映像の主観品質を推定する演算を行うことを特徴とする映像品質客観評価方法。 - 請求項5において、
前記演算ステップは、前記ビット列に含まれる動きベクトルの情報を抽出して、抽出した動きベクトルの情報から動きベクトルの統計量を演算し、
前記主観品質推定ステップは、前記予測符号化に用いられるビット量、前記変換符号化に用いられるビット量および前記符号化の制御に用いられるビット量と前記動きベクトルの統計量に基づいて前記映像の主観品質を推定する演算を行うことを特徴とする映像品質客観評価方法。 - 請求項5において、
前記演算ステップは、前記ビット列に含まれるI(イントラ)フレームまたはスライスまたはブロック、P(前方向予測)フレームまたはスライスまたはブロックおよびB(双方向、双予測)フレームまたはスライスまたはブロックの情報を抽出するとともに、抽出したI(イントラ)フレームまたはスライスまたはブロック、P(前方向予測)フレームまたはスライスまたはブロックおよびB(双方向、双予測)フレームまたはスライスまたはブロックの情報に基づいてI(イントラ)フレームまたはスライスまたはブロック、P(前方向予測)フレームまたはスライスまたはブロックおよびB(双方向、双予測)フレームまたはスライスまたはブロックの統計情報を演算し、
前記主観品質推定ステップは、前記予測符号化に用いられるビット量、前記変換符号化に用いられるビット量および前記符号化の制御に用いられるビット量、前記I(イントラ)フレームまたはスライスまたはブロック、P(前方向予測)フレームまたはスライスまたはブロックおよびB(双方向、双予測)フレームまたはスライスまたはブロックの各統計情報に基づいて前記映像の主観品質を推定する演算を行うことを特徴とする映像品質客観評価方法。 - 請求項1において、
前記演算ステップは、前記ビット列に含まれる、Iスライス、PスライスおよびBスライスのビット量と、Iスライス、PスライスおよびBスライスの量子化情報の統計量を抽出し、前記主観品質推定ステップは、前記ビット量と前記量子化情報に基づいて、前記映像の主観品質を推定する演算を行うことを特徴とする映像品質客観評価方法。 - 請求項8において、
前記ビット量とは、前記ビット列に含まれる、予測符号化に用いられる情報、変換符号化に用いられる情報および符号化の制御に用いられる情報を抽出して、抽出したこれらの情報から演算される、予測符号化に使われるビット量、変換符号化に使われるビット量および符号化の制御に使われるその他のビット量であることを特徴とする映像品質客観評価方法。 - 請求項8において、
前記ビット量とは、PスライスとBスライスの予測符号化に使われるビット量の和、変換符号化に使われるビット量の和および符号化の制御に使われるその他のビット量の和であることを特徴とする映像品質客観評価方法。 - 請求項8において、
前記量子化情報とは、PスライスとBスライスのビット列に含まれる量子化情報の平均値、もしくは和算や乗算や指数対数演算による統計量であることを特徴とする映像品質客観評価方法。 - 請求項1~11のいずれかにおいて、
Iスライスの代わりにIマクロブロック、Iフレームを使い、Pスライスの代わりにPマクロブロック、Pフレームを使い、Bスライスの代わりにBマクロブロック、Bフレームを使って前記映像の主観品質を推定することを特徴とする映像品質客観評価方法。 - 請求項1~11のいずれかにおいて、
Iモードは動き補償を用いないモード、Pモードは1つのリファレンスフレームから動き補償するモード、Bモードは2つ以上のリファレンスフレームから動き補償するモードであることを特徴とする映像品質客観評価方法。 - 映像の主観品質を評価する評価装置であって、
動き補償フレーム間予測とDCT変換もしくはウェーブレット変換等のその他の直交変換を用いて符号化された前記映像のビット列を受信する受信部と、
受信したビット列に含まれる情報を入力して所定の演算を行う第1演算部と、
前記第1演算部の演算結果に基づいて前記映像の主観品質を推定する演算を行う第2演算部と
を有することを特徴とする映像品質客観評価装置。 - 動き補償フレーム間予測とDCT変換もしくはウェーブレット変換等のその他の直交変換を用いて符号化された前記映像のビット列を受信する受信処理と、
前記受信処理に基づいて受信したビット列に含まれる情報を入力して所定の演算を行う演算処理と、
前記演算処理の演算結果に基づいて前記映像の主観品質を推定する演算を行う主観品質推定処理と
をコンピュータに実行させるプログラム。
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP09722268A EP2252073A4 (en) | 2008-03-21 | 2009-03-23 | METHOD, DEVICE AND PROGRAM FOR OBJECTIVELY EVALUATING THE QUALITY OF VIDEO CONTENT |
US12/922,846 US20110013694A1 (en) | 2008-03-21 | 2009-03-23 | Video quality objective assessment method, video quality objective assessment apparatus, and program |
CN2009801092135A CN101978700A (zh) | 2008-03-21 | 2009-03-23 | 视频质量客观评价方法、视频质量客观评价设备及程序 |
BRPI0909331A BRPI0909331A2 (pt) | 2008-03-21 | 2009-03-23 | método de avaliação objetiva de qualidade de vídeo, aparelho de vídeo, aparelho de avaliação objetiva de qualidade de vídeo, e programa |
KR20107020736A KR101188833B1 (ko) | 2008-03-21 | 2009-03-23 | 비디오 품질을 객관적으로 평가하기 위한 방법, 장치, 및 프로그램 |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2008074547 | 2008-03-21 | ||
JP2008-074547 | 2008-03-21 | ||
JP2009041457A JP2009260940A (ja) | 2008-03-21 | 2009-02-24 | 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 |
JP2009-041457 | 2009-02-24 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2009116666A1 true WO2009116666A1 (ja) | 2009-09-24 |
Family
ID=41091062
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2009/055679 WO2009116666A1 (ja) | 2008-03-21 | 2009-03-23 | 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 |
Country Status (7)
Country | Link |
---|---|
US (1) | US20110013694A1 (ja) |
EP (1) | EP2252073A4 (ja) |
JP (1) | JP2009260940A (ja) |
KR (1) | KR101188833B1 (ja) |
CN (1) | CN101978700A (ja) |
BR (1) | BRPI0909331A2 (ja) |
WO (1) | WO2009116666A1 (ja) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101895788A (zh) * | 2010-07-21 | 2010-11-24 | 深圳市融创天下科技发展有限公司 | 一种视频编码性能客观评价方法及系统 |
WO2011048829A1 (ja) * | 2009-10-22 | 2011-04-28 | 日本電信電話株式会社 | 映像品質推定装置、映像品質推定方法およびプログラム |
CN102713977A (zh) * | 2010-01-12 | 2012-10-03 | 西江大学校产学协力团 | 使用量化代码的图片质量评估方法和设备 |
JP2013527688A (ja) * | 2010-04-30 | 2013-06-27 | トムソン ライセンシング | ビデオストリームの品質を評価する方法と装置 |
Families Citing this family (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP5484140B2 (ja) * | 2010-03-17 | 2014-05-07 | Kddi株式会社 | 映像品質の客観画質評価装置 |
JP2012182785A (ja) * | 2011-02-07 | 2012-09-20 | Panasonic Corp | 映像再生装置および映像再生方法 |
EP2745518B1 (en) * | 2011-09-26 | 2017-06-14 | Telefonaktiebolaget LM Ericsson (publ) | Estimating user-perceived quality of an encoded video stream |
US9203708B2 (en) | 2011-09-26 | 2015-12-01 | Telefonaktiebolaget L M Ericsson (Publ) | Estimating user-perceived quality of an encoded stream |
KR20140096340A (ko) * | 2011-12-09 | 2014-08-05 | 텔레호낙티에볼라게트 엘엠 에릭슨(피유비엘) | 비디오 비트스트림에서 품질 결함을 검출하는 방법 및 장치 |
BR112014014349A2 (pt) * | 2011-12-15 | 2017-06-13 | Thomson Licensing | método e aparelho para medição da qualidade de vídeo |
KR101279705B1 (ko) * | 2011-12-22 | 2013-06-27 | 연세대학교 산학협력단 | 영상 프레임 내의 블러 측정 방법과 이를 이용하여 영상 프레임의 화질 측정 장치 및 방법 |
US8826314B2 (en) | 2012-01-30 | 2014-09-02 | At&T Intellectual Property I, Lp | Method and apparatus for managing quality of service |
CN103428523B (zh) * | 2012-05-22 | 2015-07-08 | 华为技术有限公司 | 评估视频质量的方法和装置 |
JP5981803B2 (ja) * | 2012-08-07 | 2016-08-31 | 日本電信電話株式会社 | 画質評価装置、画質評価方法及び画質評価プログラム |
JP5956316B2 (ja) * | 2012-11-26 | 2016-07-27 | 日本電信電話株式会社 | 主観画質推定装置、主観画質推定方法及びプログラム |
WO2014082279A1 (en) * | 2012-11-30 | 2014-06-05 | Thomson Licensing | Method and apparatus for estimating video quality |
CN103747237B (zh) * | 2013-02-06 | 2015-04-29 | 华为技术有限公司 | 视频编码质量的评估方法及设备 |
JP5992866B2 (ja) * | 2013-05-23 | 2016-09-14 | 日本電信電話株式会社 | 主観画質推定装置、及び、主観画質推定プログラム |
CN104202594B (zh) * | 2014-07-25 | 2016-04-13 | 宁波大学 | 一种基于三维小波变换的视频质量评价方法 |
CN106713901B (zh) * | 2015-11-18 | 2018-10-19 | 华为技术有限公司 | 一种视频质量评价方法及装置 |
CN105611283B (zh) * | 2015-12-24 | 2017-12-29 | 深圳市凯木金科技有限公司 | 一种主观影像效果评价方法 |
JP6275355B2 (ja) * | 2016-01-14 | 2018-02-07 | 三菱電機株式会社 | 符号化性能評価支援装置、符号化性能評価支援方法及び符号化性能評価支援プログラム |
EP3291556A1 (en) * | 2016-08-30 | 2018-03-07 | Deutsche Telekom AG | Method and apparatus for determining the perceptual video quality of a chunk of multimedia content |
JP7215229B2 (ja) | 2019-03-01 | 2023-01-31 | 日本電信電話株式会社 | 映像品質推定装置、映像品質推定方法及びプログラム |
JP7400991B2 (ja) * | 2020-10-30 | 2023-12-19 | 富士通株式会社 | サーバ、制御方法、および制御プログラム |
CN115442589B (zh) * | 2022-08-03 | 2025-01-28 | 珠海全志科技股份有限公司 | 基于视频帧信息的视频源压缩质量分析方法及装置 |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2006043500A1 (ja) * | 2004-10-18 | 2006-04-27 | Nippon Telegraph And Telephone Corporation | 映像品質客観評価装置、評価方法およびプログラム |
WO2007066066A2 (en) * | 2005-12-05 | 2007-06-14 | British Telecommunications Public Limited Company | Non-intrusive video quality measurement |
WO2009025357A1 (ja) * | 2007-08-22 | 2009-02-26 | Nippon Telegraph And Telephone Corporation | 映像品質推定装置、映像品質推定方法、フレーム種別判定方法、および記録媒体 |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5907372A (en) * | 1996-06-28 | 1999-05-25 | Hitachi, Ltd. | Decoding/displaying device for decoding/displaying coded picture data generated by high efficiency coding for interlace scanning picture format |
GB0228556D0 (en) | 2002-12-06 | 2003-01-15 | British Telecomm | Video quality measurement |
JP2006509437A (ja) * | 2002-12-10 | 2006-03-16 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | デジタルビデオ処理に対する統一測定基準(umdvp) |
-
2009
- 2009-02-24 JP JP2009041457A patent/JP2009260940A/ja not_active Withdrawn
- 2009-03-23 KR KR20107020736A patent/KR101188833B1/ko not_active Expired - Fee Related
- 2009-03-23 CN CN2009801092135A patent/CN101978700A/zh active Pending
- 2009-03-23 US US12/922,846 patent/US20110013694A1/en not_active Abandoned
- 2009-03-23 BR BRPI0909331A patent/BRPI0909331A2/pt not_active IP Right Cessation
- 2009-03-23 EP EP09722268A patent/EP2252073A4/en not_active Withdrawn
- 2009-03-23 WO PCT/JP2009/055679 patent/WO2009116666A1/ja active Application Filing
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2006043500A1 (ja) * | 2004-10-18 | 2006-04-27 | Nippon Telegraph And Telephone Corporation | 映像品質客観評価装置、評価方法およびプログラム |
WO2007066066A2 (en) * | 2005-12-05 | 2007-06-14 | British Telecommunications Public Limited Company | Non-intrusive video quality measurement |
WO2009025357A1 (ja) * | 2007-08-22 | 2009-02-26 | Nippon Telegraph And Telephone Corporation | 映像品質推定装置、映像品質推定方法、フレーム種別判定方法、および記録媒体 |
Non-Patent Citations (5)
Title |
---|
"Advanced video coding for generic audiovisual services", ITU-T H.264, February 2000 (2000-02-01) |
"Methodology for the subjective assessment of the quality of television pictures", ITU-R BT.500, 2002 |
"TELEPHONE TRANSMISSION QUALITY, TELEPHONE INSTALLATIONS, LOCAL LINE NETWORKS", ITU-T P.910, September 1999 (1999-09-01) |
D. HANDS: "Quality Assurance for IPTV", ITU-T WORKSHOP ON "END-TO-END QOE/QOS", June 2006 (2006-06-01) |
KAZUHISA YAMAGISHI; TAKANORI HAYASHI: "Video Quality Estimation Model for IPTV Services", TECHNICAL REPORT OF IEICE, July 2007 (2007-07-01), pages 123 - 126 |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2011048829A1 (ja) * | 2009-10-22 | 2011-04-28 | 日本電信電話株式会社 | 映像品質推定装置、映像品質推定方法およびプログラム |
EP2493205A1 (en) * | 2009-10-22 | 2012-08-29 | Nippon Telegraph And Telephone Corporation | Video quality estimation device, video quality estimation method, and video quality estimation program |
CN102687517A (zh) * | 2009-10-22 | 2012-09-19 | 日本电信电话株式会社 | 视频质量估计设备、视频质量估计方法以及程序 |
EP2493205A4 (en) * | 2009-10-22 | 2012-12-26 | Nippon Telegraph & Telephone | VIDEO QUALITY ASSESSMENT DEVICE, VIDEO QUALITY ASSESSMENT PROCEDURE AND VIDEO QUALITY ASSESSMENT PROGRAM |
KR101359722B1 (ko) | 2009-10-22 | 2014-02-05 | 니폰덴신뎅와 가부시키가이샤 | 비디오 품질 추정장치, 비디오 품질 추정방법 및 프로그램 |
JP5519690B2 (ja) * | 2009-10-22 | 2014-06-11 | 日本電信電話株式会社 | 映像品質推定装置、映像品質推定方法およびプログラム |
US9001897B2 (en) | 2009-10-22 | 2015-04-07 | Nippon Telegraph And Telephone Corporation | Video quality estimation apparatus, video quality estimation method, and program |
CN102713977A (zh) * | 2010-01-12 | 2012-10-03 | 西江大学校产学协力团 | 使用量化代码的图片质量评估方法和设备 |
JP2013527688A (ja) * | 2010-04-30 | 2013-06-27 | トムソン ライセンシング | ビデオストリームの品質を評価する方法と装置 |
US9288071B2 (en) | 2010-04-30 | 2016-03-15 | Thomson Licensing | Method and apparatus for assessing quality of video stream |
CN101895788A (zh) * | 2010-07-21 | 2010-11-24 | 深圳市融创天下科技发展有限公司 | 一种视频编码性能客观评价方法及系统 |
Also Published As
Publication number | Publication date |
---|---|
JP2009260940A (ja) | 2009-11-05 |
EP2252073A1 (en) | 2010-11-17 |
CN101978700A (zh) | 2011-02-16 |
US20110013694A1 (en) | 2011-01-20 |
EP2252073A4 (en) | 2011-06-08 |
KR101188833B1 (ko) | 2012-10-08 |
BRPI0909331A2 (pt) | 2015-09-29 |
KR20100126397A (ko) | 2010-12-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2009116666A1 (ja) | 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 | |
Frnda et al. | Impact of packet loss and delay variation on the quality of real-time video streaming | |
JP6328637B2 (ja) | ビデオストリーミングサービスのためのコンテンツ依存型ビデオ品質モデル | |
WO2013173909A1 (en) | Methods and apparatus for providing a presentation quality signal | |
DK2347599T3 (en) | METHOD AND SYSTEM FOR DETERMINING A QUALITY VALUE OF A VIDEO FLOW | |
JP5882320B2 (ja) | 映像信号の符号化または圧縮の間の、映像信号の品質を評価するための方法および装置 | |
Joskowicz et al. | A parametric model for perceptual video quality estimation | |
Adeyemi-Ejeye et al. | Packet loss visibility across SD, HD, 3D, and UHD video streams | |
Xu et al. | Consistent visual quality control in video coding | |
Konuk et al. | A spatiotemporal no-reference video quality assessment model | |
JP4861371B2 (ja) | 映像品質推定装置、方法、およびプログラム | |
Rossholm et al. | A new low complex reference free video quality predictor | |
JP4787303B2 (ja) | 映像品質推定装置、方法、およびプログラム | |
JP2015530807A (ja) | ビデオ品質評価のためにコンテンツ複雑性を推定する方法および装置 | |
Joskowicz et al. | A general parametric model for perceptual video quality estimation | |
Liu et al. | Real-time video quality monitoring | |
US9723266B1 (en) | Lightweight content aware bit stream video quality monitoring service | |
JP4309703B2 (ja) | 符号化誤差推定装置 | |
JP4802200B2 (ja) | 映像品質推定装置、方法、およびプログラム | |
Goudarzi | A no-reference low-complexity QoE measurement algorithm for H. 264 video transmission systems | |
Zhang et al. | Compressed-domain-based no-reference video quality assessment model considering fast motion and scene change | |
JP5523357B2 (ja) | 映像品質推定装置及び方法及びプログラム | |
Ascenso et al. | Packet-header based no-reference quality metrics for H. 264/AVC video transmission | |
JP4740967B2 (ja) | 映像品質推定装置、方法、およびプログラム | |
Karthikeyan et al. | Perceptual video quality assessment in H. 264 video coding standard using objective modeling |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 200980109213.5 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 09722268 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2009722268 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 12922846 Country of ref document: US |
|
ENP | Entry into the national phase |
Ref document number: 20107020736 Country of ref document: KR Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
ENP | Entry into the national phase |
Ref document number: PI0909331 Country of ref document: BR Kind code of ref document: A2 Effective date: 20100916 |