JP2017005344A - Video quality estimation device, video quality estimation method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To estimate the difference in video quality before and after transcode with high precision by using a video feature amount in consideration of motion compensation.SOLUTION: A video feature amount calculation device includes a difference generator for generating differential pixel information for preceding and subsequent frames of a video content before transcode, a frame feature amount calculator for calculating the feature amount of the differential pixel information, a motion compensation feature amount calculator 1 for calculating one feature amount based on the feature amount calculated for each frame of a video content, an index calculator 4 for calculating an index representing the quality difference before and after transcode of a video content, and a quality difference estimation unit 5 for calculating an estimation value of the quality difference before and after transcode based on the one feature amount and the index. The differential pixel information represents the difference of pixel information of respective pixels of the preceding and subsequent frames of the video content under the state that the pixels of only one of the frames are displaced in the opposite direction to the motion direction of a picture by the amount corresponding to the motion of the picture between the preceding and subsequent frames.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a video quality difference estimation device, a video quality difference estimation method, and a program.

  Video distribution services using network environments such as optical lines, 3G lines, LTE (Long Term Evolution), and wireless LAN (Local Area Network) have been rapidly spreading in recent years. Expected to increase rapidly. In a network environment, there is an upper limit on the amount of data that can be communicated at one time. Therefore, when a lot of traffic occurs, network line congestion occurs. When congestion occurs, communication is interrupted, and the user's satisfaction is reduced because the user feels stressed in using the communication service because the speed cannot be secured sufficiently. As one of the measures for avoiding congestion, transcoding technology for recompressing video data has been researched and developed. However, when the amount of video data is reduced by transcoding, the image quality is lowered and the quality of experience (QoE) is lowered.

  In order to provide services with comfortable QoE, it is necessary to quantify (evaluate) QoE, monitor quality fluctuations, reflect it in investment in network equipment, and control traffic. Especially in services where users create and share video content, the quality of the video before transcoding may be low. In this case, even if the video compression is reduced and the video data is reduced, Since the video quality is low, the video quality after transcoding is also low. For this reason, the quality monitoring / management for the transcoded video cannot be performed appropriately only by the video quality after the transcoding. Therefore, it is important to grasp the difference in video quality before and after transcoding in quality monitoring and management for transcoding. Specific examples of quality monitoring and management for transcoded video include monitoring whether the video quality difference before and after transcoding exceeds a certain amount, or transcoding so that the video quality difference becomes smaller than a certain amount. It is conceivable to set the compression rate of the system and operate it.

  QoE is based on a value obtained by subjectively evaluating the quality experienced by a user when viewing a video. However, in the subjective quality evaluation, the evaluator must watch the video and evaluate the quality of experience, so that the number of video contents that can be evaluated is limited due to time and money cost restrictions. For this reason, objective quality evaluation for estimating QoE from measurable physical parameters such as video signals and packet headers is used for quality management during service operation. In particular, for quality monitoring and management of transcoded video, an objective evaluation technique for estimating the difference between QoE for video before transcoding and QoE for video after transcoding is required.

  Typical examples of objective quality evaluation by comparing video signals before and after quality degradation include PSNR (Peak signal-to-noise ratio) and SSIM (Structural SIMilarity) (Non-patent Document 1). However, since the relationship between PSNR and the difference between SSIM and QoE (coefficient of regression line) depends on the video content, in order to realize the estimation of QoE difference with high accuracy using PSNR or SSIM, video Content features need to be considered. As a video feature amount that is an index representing a feature of video content, a spatial feature amount SI and a temporal feature amount TI defined in Non-Patent Document 2 are well known.

Z. Wang, AC Bovik, HR Sheikh, and EP Simoncelli, "Image quality assessment: From error visibility to structural similarity," IEEE Trans. Image Process, vol. 13, no. 4, pp. 600-612, Apr. 2004 . ITU-T Rec. P.910, "Subjective video quality assessment methods for multimedia applications," Apr. 2008.

  The variation in the difference in QoE with respect to a change in quality difference index such as SSIM depends on the difficulty of encoding for each video content (ease of quality deterioration due to encoding). In order to estimate with high accuracy, it is necessary to consider the difficulty of encoding for each video content.

  The video feature amount described in Non-Patent Document 2 can express spatial complexity and temporal complexity, and the complexity (encoding difficulty) of the video can be estimated to some extent by a combination thereof. However, in the currently popular video coding method (for example, H.264 / AVC), data is compressed by a technique called motion compensation, and SI and TI are codes of a method using this motion compensation. It is not captured difficulty.

  For example, when an image obtained by shooting a road with a fixed point camera is considered, the background of the road or the like hardly changes, and the movement of a car or a person becomes a difference between the frames. In motion compensation, a difference image is generated for a background portion with reference to the pixels in that portion of the previous and subsequent frames and compressed. For an object such as a car or a person, the direction and speed (motion vector) of the motion are searched, a difference image is generated with reference to the pixel of the movement source (or movement destination), and the difference image is compressed. SI indicates whether the background is complicated (such as whether trees with many leaves are reflected), but the compression by motion compensation refers to the previous and subsequent frames, so the background complexity does not have a significant effect. The TI represents the size, number, moving speed, etc. of the object. If the motion vector can be found correctly even if this value is large, it can be efficiently compressed by compression by motion compensation. As described above, the existing video feature quantity does not sufficiently capture the difficulty of encoding of the current encoding method.

  The present invention has been made in view of the above points, and an object of the present invention is to estimate a difference in video quality before and after transcoding with high accuracy by using a video feature amount considering motion compensation.

  Therefore, in order to solve the above-described problem, the video quality difference estimation apparatus, for the frames before and after the video content before transcoding, sets the pixels of one frame as the motion of the video as much as the video has moved between the previous and next frames. Is a difference generation unit that generates difference pixel information indicating a difference between pixel information of each pixel in the preceding and following frames in a state shifted in the reverse direction, and a frame feature amount calculation unit that calculates a feature amount of the difference pixel information; A motion compensation feature amount calculation unit that calculates one feature amount based on the feature amount calculated for each frame of the video content, and an index that calculates an index indicating a quality difference before and after transcoding of the video content Based on the calculation unit, the feature amount calculated by the motion compensation feature amount calculation unit, and the index calculated by the index calculation unit, Serial having a quality difference estimator that calculates an estimate of the quality difference between before and after transcoding.

  It is possible to estimate the difference in video quality before and after transcoding with high accuracy using the video feature amount considering the motion compensation.

It is a figure which shows the hardware structural example of the video quality difference estimation apparatus in embodiment of this invention. It is a figure which shows the function structural example of the video quality difference estimation apparatus in embodiment of this invention. It is a figure which shows the structural example of a motion compensation feature-value calculation part. It is a flowchart for demonstrating an example of the process sequence of an image quality difference estimation process. It is a flowchart for demonstrating an example of the process sequence of a motion compensation feature-value calculation process.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram illustrating a hardware configuration example of a video quality difference estimation apparatus according to an embodiment of the present invention. The video quality difference estimation device 10 in FIG. 1 includes a drive device 100, an auxiliary storage device 102, a memory device 103, a CPU 104, an interface device 105, and the like that are mutually connected by a bus B.

  A program for realizing processing in the video quality difference estimation apparatus 10 is provided by a recording medium 101 such as a CD-ROM. When the recording medium 101 storing the program is set in the drive device 100, the program is installed from the recording medium 101 to the auxiliary storage device 102 via the drive device 100. However, the program need not be installed from the recording medium 101 and may be downloaded from another computer via a network. The auxiliary storage device 102 stores the installed program and also stores necessary files and data.

  The memory device 103 reads the program from the auxiliary storage device 102 and stores it when there is an instruction to start the program. The CPU 104 executes a function related to the video quality difference estimation device 10 according to a program stored in the memory device 103. The interface device 105 is used as an interface for connecting to a network.

  FIG. 2 is a diagram illustrating a functional configuration example of the video quality difference estimation apparatus according to the embodiment of the present invention. In FIG. 2, the video quality difference estimation device 10 includes a motion compensation feature quantity calculation unit 1, a spatial feature quantity calculation unit 2, a temporal feature quantity calculation unit 3, a quality difference index calculation unit 4, and a video quality difference estimation unit 5. Have

  Each frame of the video content before transcoding is continuously input to the motion compensation feature amount calculation unit 1 according to the order. The data format of the video content is not limited to a predetermined format. The motion compensation feature amount calculation unit 1 calculates a motion compensation feature amount σ that is used as an index representing the difficulty of encoding for the video content.

  FIG. 3 is a diagram illustrating a configuration example of the motion compensation feature amount calculation unit. In FIG. 3, the motion compensation feature amount calculation unit 1 includes a block division unit 12, a difference image generation unit 13, a block integration unit 15, a frame feature amount calculation unit 16, a feature amount calculation unit 17, and the like. The motion compensation feature amount calculation unit 1 also includes a pixel information storage unit 11 and a frame feature amount storage unit 18. The pixel information storage unit 11 and the frame feature amount storage unit 18 can be realized using the memory device 103 or the auxiliary storage device 102, for example.

The pixel information storage unit 11 stores input pixel information of a frame before transcoding (hereinafter referred to as “pre-transcoding frame pixel information F _n ”), and stores the previous pre-transcoding frame pixel information (past past). Pixel information) F _n−1 is input to the difference image generation unit 13. The pixel information of the frame is, for example, the luminance signal Y, the color difference signals U and V for each pixel of the frame. Since humans are more sensitive to changes in luminance signals than changes in color difference signals, luminance signals are mainly used in the field of video quality.

The block dividing unit 12 divides the inputted pre-transcoded frame pixel information F _n into blocks, and inputs the block pixel information B _{n, i} to the difference image generating unit 13. A block is a unit formed by dividing a frame into predetermined sections. The block size is typically 16 × 16, but 8 × 8, 4 × 4, 32 × 32, or 64 × 64 may be used. Hereinafter, the block size is described assuming 16 × 16, but other sizes may be used. Note that if the number of pixels in at least one of the vertical and horizontal directions of the frame is not a multiple of 16, the size of some blocks may not be 16 × 16.

The difference image generation unit 13 is configured to change the pixels of one frame before and after the frame of the video content in a state in which the pixels of one frame are shifted in the direction opposite to the motion of the video by the amount of movement of the video between the previous and next frames. Information indicating a difference in pixel information of each pixel of the frame is generated. Here, the motion of the video in the frame is not limited to one direction. This is because when there are movements in a plurality of parts in the frame, the movement directions of the respective parts are not necessarily the same. Therefore, the difference image generation unit 13 calculates block difference pixel information DB _{n, i} for each block based on the past pixel information F _n−1 and the block pixel information B _{n, i} that are input, and each block difference Pixel information DB _{n, i} is input to the block integration unit 15. Specifically, when the pixel values at the positions x and y in the frame are F _n−1 (x, y) and the pixel values at the positions x and y in the block are represented by B _{n, i} (x, y). Then, the motion vector search unit 14 included in the difference image generation unit 13 obtains x and y that minimize the following expression (1) for each block.

式（１）が最小となるようなｘ、ｙを探すことを、動きベクトル探索という。この例では、動きベクトルの精度は１ｐｉｘｅｌであるが、フィルタを用いることで１／２ｐｉｘｅｌや１／４ｐｉｘｅｌ単位で動きベクトルを探索することもできる。

Searching for x and y that minimizes Equation (1) is called motion vector search. In this example, the accuracy of the motion vector is 1 pixel, but it is also possible to search for the motion vector in units of 1/2 pixel or 1/4 pixel by using a filter.

The difference image generation unit 13 generates the block difference pixel information DB _{n, i} for each block by executing the following substitution operation using x and y searched by the motion vector search unit 14.
DB _{n, i} (j, k) = F _n-1 (x + j, y + k) −B _{n, i} (j, k)
Here, j = 1, 2,..., 16, k = 1, 2,.
That is, “=” here means substitution. That is, the block difference pixel information DB _{n, i} is obtained by substituting the calculation result of F _n-1 (x + j, y + k) −B _{n, i} (j, k) into DB _{n, i} (j, k). Is generated. Further, as the values of x and y, values searched for the block are used. Each block difference pixel information DB _{n, i} generated is input to the block integration unit 15.

In order to prevent (x + j, y + k) from being outside the frame region, 0 ≦ x ≦ frame width (number of pixels) −16, 0 ≦ y ≦ frame width (number of pixels) −16 May be provided. Alternatively, if the range of x and y is set to a slightly larger value (such as not -16), or the frame area needs to be accessed outside the frame area by a fraction (not a multiple of 16), for example, The pixel value outside the area may be set to 0, or the value of F _n−1 (x + j, y + k) is determined by referring to the pixel value of the pixel closest to the access destination (for example, the neighbor) in the area. May be.

The block integration unit 15 integrates each input block difference pixel information DB _{n, i} to generate frame difference pixel information DF _n . Each block difference pixel information DB _n, the integration of _i, each block difference pixel information DB _{n, i} the sequence so that the relative positional relationship of each block difference pixel information DB _n, each block of the _i is maintained Then, one frame difference pixel information DF _n is generated. The generated frame difference pixel information DF _n is input to the frame feature amount calculation unit 16.

The frame feature amount calculation unit 16 calculates a frame feature amount (standard deviation of each pixel value of DF _n ) σ _n from the input frame difference pixel information DF _n . The frame feature value, in addition to the standard deviation of the frame difference pixel information DF _n, may mean square, etc. employed for the dispersion and the pixel value of each pixel value. The calculated frame feature amount σ _n is input to the feature amount calculation unit 17.

The feature amount calculation unit 17 stores the input frame feature amount σ _n in the frame feature amount storage unit 18. When the frame feature value σ _n of all the frames of the video content whose quality difference is estimated before and after transcoding is stored in the frame feature value storage unit 18, the feature value calculation unit 17 stores all the frame feature values σ Based on _n , one motion compensation feature amount σ is calculated. For example, the average value of all the frame feature values σ _n is calculated as the motion compensation feature value σ.

  As the pixel information used to calculate the motion compensation feature quantity σ, luminance information is usually used, but either a color difference signal or RGB signal may be used, and each pixel information A result obtained by weighting and summing each motion compensation feature amount calculated based on the above may be calculated as a motion compensation feature amount σ.

Returning to FIG. Each frame of video content before transcoding is continuously input to the spatial feature amount calculation unit 2 in accordance with the order. Spatial feature amount calculation unit 2, based on the transcoding previous frame pixel information F _n of each frame input to calculate the spatial feature amount. For example, the SI described in Non-Patent Document 2 may be calculated as the spatial feature amount SI. In Non-Patent Document 2, the maximum feature value for each frame is set as SI, but the average value of the feature values for each frame may be set as the spatial feature value SI.

The temporal feature amount calculation unit 3 is continuously input with each frame of video content before transcoding in the order. The temporal feature amount calculation unit 3 calculates a temporal feature amount based on the pre-transcoded frame pixel information F _n of each input frame. For example, the TI described in Non-Patent Document 2 may be calculated as the temporal feature amount TI. In Non-Patent Document 2, the maximum value of the feature value for each frame is TI, but the average value of the feature values for each frame may be the temporal feature value TI.

Each frame of video content before transcoding and each frame of video content after transcoding are continuously input to the quality difference index calculation unit 4 in the order. The quality difference index calculation unit 4 calculates a quality difference index based on the pre-transcoded frame pixel information F _{n of} each input frame and the post-transcoded frame pixel information F ′ _{n of} each input frame. . For example, SSIM may be calculated as a quality difference index. Alternatively, PSNR may be used as a quality difference index, and a combination of SSIM and PSNR may be calculated as a quality difference index.

The video quality difference estimation unit 5 determines the quality difference before and after transcoding of video content based on the input motion compensation feature quantity σ, spatial feature quantity SI, temporal feature quantity TI, and quality difference index SSIM. A video quality difference DMOSp that is an estimated value of is calculated. For example, the video quality difference DMOSp may be calculated by the following equation (2).
DMOSp = (c ₁ · σ + c ₂ · SI + c ₃ · TI + c ₄ ) · (1-SSIM) (2)
Here, c _{1 to} c ₄ are constants, and can be freely determined by the designer of the video quality difference estimation apparatus 10 based on data obtained through subjective quality evaluation experiments.

  Hereinafter, a processing procedure executed by the video quality difference estimation device 10 will be described. FIG. 4 is a flowchart for explaining an example of the processing procedure of the video quality difference estimation processing.

  In step S <b> 1-1, the motion compensation feature quantity calculation unit 1 calculates a motion compensation feature quantity σ for video content for which a video quality difference is to be estimated (hereinafter referred to as “target content”).

In addition, the spatial feature amount calculation unit 2 starts a processing procedure for calculating the spatial feature amount SI in response to the input of the pre-transcoding frame pixel information F _n and outputs the calculated spatial feature amount SI. (Step S1-2).

Further, the temporal feature amount calculation unit 3 starts the processing procedure for calculating the temporal feature amount TI in accordance with the input of the transcoding previous frame pixel information F _n, it outputs the temporal feature amount TI calculated (Step S1-3).

Further, the quality difference index calculation unit 4 outputs the video quality difference index SSIM in response to the input of the pre-transcoding frame pixel information F _n and the post-transcoding frame pixel information F ′ _n (step S1-4).

  Subsequently, the video quality difference estimation unit 5 includes the motion compensation feature amount σ calculated by the motion compensation feature amount calculation unit 1, the spatial feature amount SI calculated by the spatial feature amount calculation unit 2, and the temporal feature. The video quality difference DMOSp is estimated based on the temporal feature quantity TI calculated by the quantity calculation unit 3 and the video quality difference index SSIM calculated by the quality difference index calculation unit 4 (step S2).

  Note that the video quality difference DMOSp may be estimated without using the spatial feature quantity SI and the temporal feature quantity TI.

  Next, details of step S1-1 will be described. FIG. 5 is a flowchart for explaining an example of the processing procedure of the motion compensation feature amount calculation processing. The motion compensation feature amount calculation unit 1 starts the processing procedure of FIG. 5 in response to the input of the first pre-transcoded frame pixel information of the target content.

When the first transcoding previous frame pixel information F ₁ is input, the motion compensation feature amount calculating unit 1 initializes the variable n for counting the number of frames (step S10). Specifically, 0 is assigned to the variable n. Subsequently, the motion compensation feature amount calculation unit 1 increments the variable n (adds 1 to the variable n), and counts the number of frames (step S11). The pixel information storage unit 11 stores the transcoded previous frame pixel information F ₁ inputted (step S12).

Subsequently, the block dividing unit 12 determines whether or not the value of the variable n is 1 (step S13). If the variable n is 1 (Yes in step S13), the process returns to S11. In this case, when the following transcoding previous frame pixel information F ₂ is input, step S11 and subsequent steps are executed.

When the variable n is not 1 (No in Step S13), the block dividing unit 12 divides the pre-transcoding frame pixel information _Fn into blocks, and generates block pixel information _{Bn, i} (Step S14).

Subsequently, the motion vector search unit 14 searches the motion vector of the block pixel information B _{n, i} for each block with reference to the pre-transcoding frame pixel information F _n . The difference image generation unit 13 generates block difference image information DB _{n, i} for each block based on the motion vector search result (step S15).

Subsequently, the block integration unit 15 integrates the block difference pixel information DB _{n, i} for each block to generate frame difference pixel information DF _n (step S16).

Subsequently, the frame feature amount calculation unit 16 calculates a frame feature amount σ _n from the frame difference pixel information DF _n (step S17). The calculated frame feature quantity σ _n is stored in the frame feature quantity storage unit 18 (step S18).

  If all the frames of the target content have been processed, the process proceeds to step S20. If not, step S11 and subsequent steps are repeated (step S19).

In step S20, the feature amount calculation unit 17 calculates a motion compensation feature amount σ based on the frame feature amount σ _n of each frame stored in the frame feature amount storage unit 18 (step S20).

Next, the effect of this embodiment will be described. The correlation coefficient (PCC) between the QoE difference (DMOS) obtained by the subjective quality evaluation test for a certain video group before and after transcoding and the estimation result (DMOSp) according to the present embodiment is 0.86. And the root mean square error (RMSE) was 0.19. When the motion compensation feature quantity σ which is a part of the feature of the present embodiment is not used (when c ₁ is 0 in the equation (2)), the PCC is 0.79, and the RMSE is 0.26. It was confirmed that the quality difference can be estimated more accurately when the motion compensation feature quantity σ is used.

  As described above, according to the present embodiment, it is possible to calculate the feature amount representing the difficulty of video coding based on the viewpoint of QoE by considering the motion compensation used in video coding. The difference in video quality before and after transcoding can be estimated with high accuracy.

  In the present embodiment, the difference image generation unit 13 is an example of a difference generation unit. The quality difference index calculation unit 4 is an example of an index calculation unit. The video quality difference estimation unit 5 is an example of a quality difference estimation unit.

  As mentioned above, although the Example of this invention was explained in full detail, this invention is not limited to such specific embodiment, In the range of the summary of this invention described in the claim, various deformation | transformation・ Change is possible.

DESCRIPTION OF SYMBOLS 1 Motion compensation feature-value calculation part 2 Spatial feature-value calculation part 3 Temporal feature-value calculation part 4 Quality difference index calculation part 5 Video quality difference estimation part 10 Video quality difference estimation apparatus 11 Pixel information storage part 12 Block division part 13 Difference Image generation unit 14 Motion vector search unit 15 Block integration unit 16 Frame feature amount calculation unit 17 Feature amount calculation unit 18 Frame feature amount storage unit 100 Drive device 101 Recording medium 102 Auxiliary storage device 103 Memory device 104 CPU
105 Interface device B bus

Claims (8)

For the frames before and after the video content before transcoding, each of the previous and next frames in a state where the pixels of one frame are shifted in the direction opposite to the movement of the video by the amount of movement of the video between the previous and next frames. A difference generation unit that generates difference pixel information indicating a difference in pixel information of pixels;
A frame feature value calculation unit for calculating a feature value of the difference pixel information;
A motion compensation feature amount calculation unit that calculates one feature amount based on the feature amount calculated for each frame of the video content;
An index calculation unit for calculating an index indicating a quality difference before and after transcoding of the video content;
A quality difference estimation unit that calculates an estimated value of the quality difference before and after the transcoding based on the feature amount calculated by the motion compensation feature amount calculation unit and the index calculated by the index calculation unit;
A video quality difference estimation device characterized by comprising: The difference generation unit includes a motion vector search unit that searches for a motion vector between the preceding and succeeding frames for each block formed by dividing the frame into predetermined sections. Generating difference pixel information indicating a difference in pixel information of each pixel of the preceding and succeeding frames in a state in which the pixel is shifted by a motion vector related to the block,
The frame feature amount calculation unit calculates a feature amount for a result of integrating the difference pixel information for each block;
The video quality difference estimation apparatus according to claim 1, wherein: A spatial feature amount calculation unit that calculates a spatial feature amount based on pixel information of a frame of the video content before transcoding;
A temporal feature amount calculation unit that calculates a temporal feature amount based on pixel information of the frame of the video content before transcoding,
The quality difference estimation unit is based on the feature amount calculated by the motion compensation feature amount calculation unit, the index calculated by the index calculation unit, the spatial feature amount, and the temporal feature amount. Calculating an estimated value of the quality difference before and after the transcoding,
The video quality difference estimation apparatus according to claim 1 or 2, wherein The index calculation unit calculates an index indicating a quality difference based on frame pixel information before and after transcoding.
The video quality difference estimation apparatus according to any one of claims 1 to 3, wherein Computer
For the frames before and after the video content before transcoding, each of the previous and next frames in a state where the pixels of one frame are shifted in the direction opposite to the movement of the video by the amount of movement of the video between the previous and next frames. A difference generation procedure for generating difference pixel information indicating a difference in pixel information of pixels;
A frame feature amount calculation procedure for calculating a feature amount of the difference pixel information;
A motion compensation feature amount calculation procedure for calculating one feature amount based on the feature amount calculated for each frame of the video content;
An index calculation procedure for calculating an index indicating a quality difference before and after transcoding of the video content;
A quality difference estimation procedure for calculating an estimated value of the quality difference before and after the transcoding based on the feature amount calculated in the motion compensation feature amount calculation procedure and the index calculated in the index calculation procedure;
A video quality difference estimation method characterized by executing: The difference generation procedure includes a motion vector search unit that searches for a motion vector between the preceding and succeeding frames for each block formed by dividing the frame into predetermined sections. Generating difference pixel information indicating a difference in pixel information of each pixel of the preceding and succeeding frames in a state in which the pixel is shifted by a motion vector related to the block,
The frame feature amount calculation procedure calculates a feature amount for a result of integrating the difference pixel information for each block,
6. The video quality difference estimation method according to claim 5, wherein: The computer is
A spatial feature amount calculating procedure for calculating a spatial feature amount based on pixel information of the frame of the video content before transcoding;
Performing a temporal feature amount calculation procedure for calculating a temporal feature amount based on pixel information of the frame of the video content before transcoding,
The quality difference estimation procedure is based on the feature amount calculated in the motion compensation feature amount calculation procedure, the index calculated in the indicator calculation procedure, the spatial feature amount, and the temporal feature amount. Calculating an estimated value of the quality difference before and after the transcoding,
7. The video quality difference estimation method according to claim 5, wherein the video quality difference is estimated.   The program for functioning a computer as each part as described in any one of Claims 1 thru | or 4.

Images