JP6431449B2 - Video matching apparatus, video matching method, and program - Google Patents

Description

  The present invention relates to a video matching device, a video matching method, and a program.

  In order to confirm that the video service is provided with good quality, the quality of the video experienced by the user is measured before or during service provision, and the quality of the video provided to the user is good. It is important to confirm and monitor this. Therefore, a video quality evaluation technique that can appropriately evaluate the video quality experienced by the user is necessary.

  Methods for evaluating video quality include a subjective evaluation method (for example, see Non-Patent Document 1) and an objective evaluation method (for example, see Non-Patent Document 2). The subjective evaluation method requires a large number of users to evaluate images in a facility where the evaluation environment (such as room illuminance and room noise) can be reproduced. For this reason, the cost is high and the evaluation takes time, so that it is not suitable for use in which quality is evaluated in real time. Therefore, development of an objective video quality evaluation method that derives physical feature quantities from video signals and outputs video quality evaluation values is underway (see Non-Patent Document 2, for example).

  Non-Patent Document 2 discloses an objective evaluation method for estimating video quality by comparing physical feature amounts of a reference video signal and a degraded video signal that has deteriorated due to encoding or network transmission of the reference video signal. It has been described. In such a method, the spatial and temporal positions of the reference video signal and the degraded video signal need to be matched. That is, it is necessary that the time-direction deviation and the spatial position deviation are consistent between the reference video signal and the degraded video signal. In Patent Document 1, in consideration of video distribution over the Internet, spatial and temporal positional alignment when the reference video signal and the degraded video signal are out of synchronization due to fluctuations in packet arrival intervals and occurrence of packet loss. A method has been proposed.

Japanese Patent No. 5347012

ITU-T recommendation P.910 "Subjective video quality assessment methods for multimedia applications" ITU-T recommendation J.247 "Objective perceptual multimedia video quality measurement in the presence of a full reference" Business Overview of Next Generation Broadcast Promotion Forum http://www.nextv-f.jp/pdf/press20130617-2.pdf ARIB STD-B56 version 1.1, [online], [Search June 9, 2015], Internet (URL: http://www.arib.or.jp/english/html/overview/doc/2-STD- B56v1_1.pdf) THE World Heritage 4K, [online], [Search June 9, 2015], Internet (URL: http://www.bs-tbs.co.jp/sekaiisan4ksetsumei/) Hikari TV 4K, [online], [Search June 9, 2015], Internet (URL: http://www.hikaritv.net/4k/)

  On the other hand, in recent years, attention has been paid to 4K / 8K video services with higher resolution than conventional full HD video (for example, see Non-Patent Document 3). In the 4K / 8K video service, the frame rate is increased from the conventional 30 fps to 60 fps or 120 fps (see, for example, Non-Patent Document 4), and the change in the luminance value before and after the frame is small. In the 4K / 8K video service, in order to emphasize high resolution, production of a program with less video motion such as a still image is in progress (see, for example, Non-Patent Documents 5 and 6).

  Therefore, unlike the packet fluctuation or packet loss phenomenon that is the subject of the prior art such as Patent Document 1, the change in the luminance value before and after the frame is small, and the temporal change is small. There is a problem that it is difficult to take.

  The present invention has been made in view of the above points, and an object of the present invention is to improve the accuracy of temporal position matching with a deteriorated video signal even for a video signal having a small temporal change.

Therefore, in order to solve the above-described problem, the video matching device includes a deriving unit that derives respective feature amounts of the first video signal and the second video signal in which the video quality of the first video signal has changed , A classification for classifying the first video signal into one of a plurality of types classified based on at least the magnitude of temporal change of the video based on the feature quantity of the first video signal. And the first video signal and the second video signal based on the respective feature amounts of the first video signal and the second video signal corresponding to the type into which the first video signal is classified. And a matching unit for calculating a temporal positional shift amount from the video signal.

  Even for a video signal having a small temporal change, it is possible to improve the accuracy of temporal position matching with a deteriorated video signal. Therefore, even in the 4K / 8K video service, an objective evaluation method for accurately estimating the video quality can be implemented.

It is a figure which shows the hardware structural example of the image | video matching apparatus in embodiment of this invention. It is a figure which shows the function structural example of the image | video matching apparatus in embodiment of this invention. It is a figure for demonstrating the 1st classification method to A group or other than that. It is a figure for demonstrating the 2nd classification method to A group or other than that. It is a figure for demonstrating the 1st classification method to B group or C group. It is a figure for demonstrating the 2nd classification method to B group or C group. It is a figure for demonstrating the 3rd classification method to B group or C group. It is a figure for demonstrating the calculation method of frame deviation | shift amount (tau) of A group. It is a figure which shows the matching reference | standard video signal R2 and the matching degradation video signal P2. It is a figure for demonstrating the calculation method of frame shift amount (tau) of B group and C group.

  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 matching device according to an embodiment of the present invention. The video matching 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.

  A program for realizing processing in the video matching 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 functions related to the video matching device 10 in accordance with 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 matching apparatus according to the embodiment of the present invention. In FIG. 2, the video matching apparatus 10 includes a video feature quantity deriving unit 11, a video classifying unit 12, a video matching unit 13, and the like. Each of these units is realized by processing executed by the CPU 104 by one or more programs installed in the video matching apparatus 10. The video matching apparatus 10 also uses the video classification information storage unit 14. The video classification information storage unit 14 can be realized by using, for example, a storage device that can be connected to the auxiliary storage device 102 or the video matching device 10 via a network.

  The input data to the video matching device 10 is the reference video signal R1 and the degraded video signal P1. The degraded video signal P1 refers to a video signal whose video quality has deteriorated as a result of the reference video signal R1 undergoing encoding, network transmission, or the like. The video matching device 10 is used, for example, when evaluating the quality of a video signal that has deteriorated due to encoding, network transmission, or the like. Used when aligning positions.

  The video feature quantity deriving unit 11 derives a feature quantity for each of the input reference video signal R1 and the degraded video signal P1. Specifically, the amount of spatial information (SI: Spatial perceptual information), time information (TI: Temporal perceptual information) specified in ITU-TP.910 shown in Non-Patent Document 1, and PSNR representing the encoding difficulty level (PeakSignal-to-NoiseRatio) is derived. The amount of spatial information and the amount of temporal information are derived for each frame of each input video signal. Hereinafter, the spatial information amount derived from the reference video signal R1 is represented as SI_R1 (k) (k = 1, 2,..., R), and the temporal information amount derived from the reference video signal R1 is expressed as TI_R1 (k ) (k = 1,2, ..., r). Further, the spatial information amount derived from the degraded video signal P1 is expressed as SI_P1 (k) (k = 1, 2,..., P), and the temporal information amount derived from the degraded video signal P1 is expressed as TI_P1 (k ) (k = 1,2, ..., p).

  Here, r is the number of frames of the reference video signal R1, and p is the number of frames of the degraded video signal P1. The number of frames of the reference video signal R1 and the number of frames of the degraded video signal P1 are not necessarily the same because frames are lost due to encoding, network transmission, etc., or the number of frames increases due to noise, etc. This is because there is a possibility of doing.

  The PSNR is derived by comparing a video obtained by coding the reference video signal R1 at n types of coding bit rates BR (l) (1 ≦ l ≦ n) and the reference video signal R1. The derived result is expressed as PSNR_R1 (l) (1 ≦ l ≦ n). PSNR_R1 (l) (1 ≦ l ≦ n) is an average PSNR for r frames when the reference video signal R1 is encoded at a certain encoding bit rate BR (l). The codec used for encoding is a comparison between PSNR_R1 (l) (1 ≦ l ≦ n) and the encoding difficulty level classification diagram (FIG. 7) stored in the video classification information storage unit 14. In order to make it possible, it is desirable that the codec is the same as the codec used when creating the encoding difficulty level classification diagram. For example, standard software that is studied in international standardization may be used as the codec.

  The video classification unit 12 includes a spatial information amount (SI_R1), a temporal information amount (TI_R1), or an encoding difficulty (PSNR_R1) of the reference video signal R1 output from the video feature quantity deriving unit 11, and a video classification information storage unit 14, the reference video signal R1 is classified into one of the three types, and the video classification result X is output. Specifically, first, the video classification unit 12 determines, based on the temporal information amount TI value, that the reference video signal R1 includes a video group type (group A) that includes a large temporal change and a video group type that does not. And classified. When the reference video signal R1 does not correspond to the group A, the video classification unit 12 converts the reference video signal R1 into a complex video group type (based on the spatial information SI value and the PSNR indicating the encoding difficulty ( Group B) or a type of video group that is not (Group C).

  Here, the group A is a video group including a large temporal change. Specifically, a video including a scene change is included in the A group. Group B is a complex video group that does not include large temporal changes, has a wide distribution of pixel values within a frame, and is a complex video group. Specifically, group B does not include scene changes, but is a complex video in which objects of various brightness exist in the frame. For example, a person or object with a motion in the background with a little motion It is an image that is mixed. Group C is a video group that does not include large temporal changes and has a narrow distribution of pixel values within a frame. Specifically, the group C is an image in which only an object that hardly moves is present, such as a forest or blue sky image that does not include a scene change. The reason for classifying the video is that the video matching unit 13 derives the frame shift amount τ between the reference video signal R1 and the degraded video signal P1 with a smaller calculation amount.

  The video classification information storage unit 14 stores information used for classification of the reference video signal R1. For example, the video classification information storage unit 14 stores a threshold value Th_TI, a threshold value Th_SI, a threshold value Th_PSNR (m), an encoding difficulty level classification diagram, a threshold value Th_std_PSNR, and the like. The threshold value Th_TI is a threshold value for the TI value of the reference video signal R1 for classifying the A group and other than the A group. The threshold value Th_SI, the threshold value Th_PSNR (m), the encoding difficulty level classification diagram, and the threshold value Th_std_PSNR are the SI value of the reference video signal R1 or the threshold value for the PSNR for classifying the B group and the C group.

  The video matching unit 13 calculates a frame shift amount τ based on the feature amount of the reference video signal R1 and the feature amount of the degraded video signal P1 by a method according to the video classification result X output from the video classification unit 12. .

  Hereinafter, the processing procedure executed by the video classification unit 12 and the video matching unit 13 will be described in more detail. FIG. 3 is a diagram for explaining a first classification method for group A or other groups.

  In step S101, the video classification unit 12 determines that the maximum value (max (TI_R1)) of the TI values (TI_R1 (k) (k = 1, 2,..., R)) of the reference video signal R1 is the video classification information storage unit. It is determined whether or not the threshold value Th_TI stored in 14 is exceeded.

  When max (TI_R1) exceeds the threshold Th_TI (Yes in step S101), the video classification unit 12 classifies the reference video signal R1 into the A group (step S102). Here, when max (TI_R1) is larger than Th_TI, it means that a scene change is included. Based on the time (t_TI_max_R1) at which the TI value becomes maximum, the reference video signal R1 and the degraded video signal P1 can be easily obtained. It is possible to match the temporal position with

  On the other hand, when max (TI_R1) is equal to or less than the threshold value Th_TI (No in step S101), the video classification unit 12 executes processing for classifying the reference video signal R1 into the B group or the C group (step S103). .

  Instead of using max (TI_R1), as shown in FIG. 4, the differential value (TI_R1) of the TI value (TI_R1 (k) (k = 1, 2,..., R)) of the reference video signal R1. )), And the reference video signal R1 is set to A group or depending on whether or not the maximum value max (TI_R1 ') of TI_R1' exceeds a threshold (Th_TI ') stored in the video classification information storage unit 14. It may be classified other than that. Here, a large differential value of the TI value means a scene change timing.

  Next, details of step S103 (classification process into the B group or the C group) will be described. In step S103, the reference video signal R1 is classified into the B group or the C group based on the SI value or PSNR of the reference video signal R1. The reason for classifying into the B group or the C group based on the SI value and PSNR is that a large SI value means that the pixel value distribution in the frame is wide, and that it is a complex video in which various objects exist. In other words, the low PSNR means that the video is complicated and difficult to encode.

  Here, three classification methods from the first method to the third method will be described. In the first method, based on the maximum value (max (SI_R1)) of the SI values (SI_R1 (k) (k = 1, 2,..., R)) of the reference video signal R1, the reference video signal R1 is grouped into the B group. Or it is the method of classifying into C group. FIG. 5 is a diagram for explaining a first classification method into the B group or the C group.

  In step S201, the video classification unit 12 determines that the maximum value (max (SI_R1)) of the SI value (SI_R1 (k) (k = 1, 2,..., R)) of the reference video signal R1 is the video classification information storage unit. It is determined whether or not the threshold value Th_SI stored in 14 is exceeded. When max (SI_R1) exceeds the threshold value Th_SI (Yes in step S202), the video classification unit 12 classifies the reference video signal R1 into the B group (step S202). Here, a large SI value means that the pixel value distribution in the frame is wide and the image is a complex image in which various objects exist.

  On the other hand, when max (SI_R1) is equal to or smaller than the threshold Th_SI (No in step S202), the video classification unit 12 classifies the reference video signal R1 into the group C (step S203).

  Next, the second method will be described. In the second method, based on PSNR (PSNR_R1 (m)) when the reference video signal R1 is encoded at one encoding bit rate (BR (m)), the reference video signal R1 is group B or group C. It is a method to classify. FIG. 6 is a diagram for explaining a second classification method into the B group or the C group.

  In step S301, the video classification unit 12 stores the PSNR (PSNR_R1 (m)) when the reference video signal R1 is encoded at one encoding bit rate BR (m) in the video classification information storage unit 14. It is determined whether it is less than a certain threshold Th_PSNR (m). When PSNR_R1 (m) is less than the threshold Th_PSNR (m) (Yes in step S301), the video classification unit 12 classifies the reference video signal R1 into the group B (step S302). Here, PSNR_R1 (m) being less than the threshold Th_PSNR (m) means that the video is complicated and difficult to encode.

  On the other hand, when PSNR_R1 (m) is greater than or equal to the threshold Th_PSNR (m) (No in step S301), the video classification unit 12 classifies the reference video signal R1 into the C group (step S303).

  Subsequently, the third method will be described. The third method classifies the reference video signal R1 into the B group or the C group based on the characteristics of each PSNR (PSNR_R1 (l)) when the reference video signal R1 is encoded with n encoding bit rates. It is a method to do. FIG. 7 is a diagram for explaining a third classification method into the B group or the C group.

  The video classification unit 12 uses the PSNR (PSNR_R1 (l)) obtained when the reference video signal R1 is encoded at n encoding bit rates, and the encoding difficulty level stored in the video classification information storage unit 14 Based on the classification diagram (FIG. 7), classification is made into a group having a closer characteristic. Group B is a video group in which PSNR changes greatly due to encoding, and is a complex video group that is difficult to encode. On the other hand, the group C is a video group in which a change in PSNR due to encoding is small, and is a video group that can be encoded easily. According to the encoding difficulty level classification diagram (FIG. 7), the PSNR of the video belonging to the B group varies greatly according to the encoding bit rate, and the PSNR of the video belonging to the C group depends on the encoding bit rate. It can be seen that the change is small. Thus, based on the standard deviation of PSNR (std_PSNR_R1) calculated using the following equation (1), when the standard deviation of PSNR (std_PSNR_R1) exceeds the threshold Th_std_PSNR, the reference video signal R1 is classified into the B group. When the standard deviation of PSNR (std_PSNR_R1) is less than or equal to the threshold value Th_std_PSNR, the reference video signal R1 may be classified into the C group.

次に、映像整合部１３が実行する処理について説明する。まず、Ａ群に関してフレームずれ量τを算出する方法を述べる。図８は、Ａ群のフレームずれ量τの算出方法を説明するための図である。

Next, processing executed by the video matching unit 13 will be described. First, a method for calculating the frame shift amount τ for the A group will be described. FIG. 8 is a diagram for explaining a method of calculating the frame shift amount τ of the A group.

  Since the video belonging to the A group includes a scene change, the video matching unit 13 matches the temporal position of the reference video signal R1 and the degraded video signal P1 at the timing of the scene change. Specifically, the video matching unit 13 sets TI values (TI_R1 (k) (k = 1, 2,..., R), TI_P1 (k) (k =) of the reference video signal R1 and the degraded video signal P1. Deriving a frame (t_TImax_R1, t_TImax_P1) that maximizes 1,2, ..., p)). The video matching unit 13 calculates the frame shift amount τ between the reference video signal R1 and the deteriorated video signal P1 based on t_TImax_R1 and t_TImax_P1, using the following equation (2).

τ = t_TImax_R1−t_TImax_P1 (2)
From this, it can be seen that the deteriorated video signal P1 and the reference video signal R1 are shifted by τ frames. Note that the frame shift amount τ may be calculated from the timing at which the differential values of TI_R1 and TI_P1 are maximized.

  In the case of τ> 0, it means that the degraded video signal P1 is advanced by τ frames from the reference video signal R1, and therefore the video matching unit 13 displays the degraded video signal P1 as shown in FIG. The matching deteriorated video signal P2 excluding the τ frame from the beginning and the matching reference video signal R2 excluding the τ frame from the end of the reference video signal R1 are output.

  When τ <0, it means that the reference video signal R1 is ahead of the degraded video signal P1 by τ frames. Therefore, the video matching unit 13 excludes the τ frame from the beginning of the reference video signal R1. The reference video signal R2 and the matched and degraded video signal P2 excluding the τ frame from the end of the degraded video signal P1 are output.

  Next, a method for calculating the frame shift amount τ for the B group and the C group will be described. Along with higher resolution and higher definition such as 4K / 8K video, the frame rate is improved. As the frame rate increases, the change between TI values becomes smaller because the change between frames becomes smaller. However, even if the change between frames is small, it is not a still image, so the luminance value in the frame changes. Therefore, the video matching unit 13 derives a frame shift amount τ between the reference video signal R1 and the degraded video signal P1 based on the SI value.

  FIG. 10 is a diagram for explaining a method of calculating the frame shift amount τ of the B group and the C group. As shown in FIG. 10, the video matching unit 13 performs the SI value (SI_R1 (k) (k = 1, 2,..., R)) of the reference video signal R1 and the SI value (SI_P1 ( k = 1, 2,..., p)), and a frame shift amount τ that maximizes the cross correlation coefficient is derived.

  Note that the number of frames to be used in deriving the cross-correlation coefficient is changed between the B group and the C group in consideration of the fact that the video complexity differs between the B group and the C group. Specifically, if the number of frames to be analyzed for Group B is Frame_B and the number of frames to be analyzed for Group C is Frame_C, then Frame_B <Frame_C. Note that Frame_C is a value that matches the smaller number of frames of the number of frames of the reference video signal R1 and the number of frames of the degraded video signal P1.

  As described above, the reason for changing the number of frames when deriving the cross-correlation coefficient between the B group and the C group is that the change in the SI value is larger in the B group than in the C group. This is because the frame shift amount τ can be calculated from the cross-correlation coefficient. Since the SI value of Group C is small, it is difficult to capture the change in a short time. However, by obtaining the cross-correlation coefficient with a certain amount of time, it is possible to capture a slight change, and the reference image It is possible to derive the frame shift amount τ between the signal R1 and the deteriorated video signal P1. Regarding the group B, the number of frames to be analyzed can be set to Frame_C, but the amount of calculation can be reduced by using Frame_B having a smaller number of frames.

  Regarding the frame shift amount τ between the reference video signal R1 derived from the cross-correlation coefficient and the degraded video signal P1, if τ> 0, the degraded video signal P1 is advanced by τ frames from the reference video signal R1. Therefore, the video matching unit 13 outputs a matched degraded video signal P2 excluding the τ frame from the beginning of the degraded video signal P1 and a matched reference video signal R2 from which the τ frame is excluded from the end of the reference video signal R1. To do. When τ <0, it means that the reference video signal R1 is ahead of the degraded video signal P1 by τ frames. Therefore, the video matching unit 13 excludes the τ frame from the beginning of the reference video signal R1. The reference video signal R2 and the matched and degraded video signal P2 excluding the τ frame from the end of the degraded video signal P1 are output.

  Alternatively, the frame position of the reference video signal R1 and the frame position of the degraded video signal P1 for matching may be output instead of the matched degraded video signal P2 and the matched reference video signal R2 itself.

  Through the above processing, the matching reference video signal R2 and the matching deteriorated video signal P2 are output. Alternatively, the frame position of the reference video signal and the frame position of the degraded video signal for matching are output, and the matched reference video signal R2 and the matched degraded video signal P2 are calculated using these. By using the matching reference video signal R2 and the matching deteriorated video signal P2, it is possible to perform appropriate quality estimation in a state where the reference video signal R1 and the deteriorated video signal P1 are temporally matched.

  As described above, according to the present embodiment, when comparing the physical feature values of the reference video signal and the deteriorated video signal in order to estimate the subjective quality, even for a video with a small temporal change. Thus, the accuracy of the temporal position alignment can be improved.

  In the present embodiment, the video feature quantity deriving unit 11 is an example of a deriving unit. The video classification unit 12 is an example of a classification unit. The video matching unit 13 is an example of a matching unit. The threshold value Th_TI or the threshold value Th_TI ′ is an example of a first threshold value. The threshold value Th_SI is an example of a second threshold value. The threshold value Th_R1 (m) is an example of a third threshold value.

  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 10 Image | video matching apparatus 11 Image | video feature-value derivation | leading-out part 12 Image | video classification | category part 13 Image | video matching part 14 Image | video classification information storage part 100 Drive apparatus 101 Recording medium 102 Auxiliary storage apparatus 103 Memory apparatus 104 CPU
105 Interface device
R1 reference video signal
P1 Degraded video signal
r Number of frames of reference video signal R1
p Number of frames of degraded video signal P1
SI_R1 (k) Spatial information amount of the kth frame of the reference video signal R1
TI_R1 (k) Time information amount of the kth frame of the reference video signal R1
PSNR_R1 (k) Encoding difficulty of the kth frame of the reference video signal R1
SI_P1 (k) Spatial information amount of frame k of degraded video signal P1
TI_P1 (k) Amount of time information in frame k of degraded video signal P1
X Video classification results A, B, C Video classification type
Th_TI TI value threshold
max (TI_R1) Maximum value of TI_R1
TI_R1 'Differential value of TI_R1
Threshold for differential value of Th_TI'TI value
max (SI_R1) Maximum value of SI_R1
Th_SI SI value threshold
BR encoding bit rate
n Types of encoding bit rates
PSNR_R1 (l) PSNR when the reference video signal R1 is encoded at the encoding bit rate BR (l)
Th_PSNR (m) PSNR threshold
std_PSNR_R1 Standard deviation of PSNR_R1 (l)
Th_std_PSNR PSNR standard deviation threshold
t_TImax_R1 Frame that maximizes the TI value of the reference video signal R1
t_TImax_P1 Frame where the TI value of the degraded video signal P1 is maximum τ Frame shift amount between the reference video signal R1 and the degraded video signal P1
Frame_B Time length for calculating the cross-correlation coefficient of video belonging to group B
Frame_C Time length for calculating the cross-correlation coefficient of video belonging to group C
R2 alignment reference video signal
P2 alignment degraded video signal

Claims (8)

A derivation unit for deriving respective feature amounts of the first video signal and the second video signal in which the video quality of the first video signal has changed ;
A classification for classifying the first video signal into one of a plurality of types classified based on at least the magnitude of temporal change of the video based on the feature quantity of the first video signal. And
The first video signal and the second video signal based on the respective feature quantities of the first video signal and the second video signal corresponding to the classified type of the first video signal. A matching unit that calculates the amount of positional deviation with respect to
An image matching apparatus comprising: The derivation unit derives a time information amount for each frame of the first video signal and the second video signal,
The classification unit classifies the first video signal by comparing a time information amount for each frame of the first video signal with a first threshold value,
When the amount of time information for each frame of the first video signal is greater than the first threshold, the matching unit includes the first video signal frame, the second video signal frame , To calculate the amount of time positional deviation of
The image matching apparatus according to claim 1, wherein: The derivation unit further derives a spatial information amount for each frame of the first video signal and the second video signal,
When the time information amount for each frame of the first video signal is equal to or less than the first threshold, the matching unit determines the spatial information amount for each frame of the first video signal and the second video. Deriving a cross-correlation coefficient with the amount of spatial information for each frame of the signal, and deriving a temporal position shift amount of the frame related to the cross-correlation coefficient.
The image matching apparatus according to claim 2, wherein: The classifying unit, when a time information amount for each frame of the first video signal is equal to or less than the first threshold, a maximum value of a spatial information amount for each frame of the first video signal; Comparing the threshold of 2 to classify the first video signal;
When the amount of time information for each frame of the first video signal is equal to or less than the first threshold, the matching unit performs the mutual operation according to the classification result of the first video signal by the classification unit. Change the number of frames used to derive the correlation coefficient,
The image matching apparatus according to claim 3, wherein: The derivation unit derives an encoding difficulty level when the first video signal is encoded at one encoding bit rate,
The classifying unit compares the encoding difficulty level of the first video signal with a third threshold when the amount of time information for each frame of the first video signal is equal to or less than the first threshold. And classifying the first video signal,
When the amount of time information for each frame of the first video signal is equal to or less than the first threshold, the matching unit performs the mutual operation according to the classification result of the first video signal by the classification unit. Change the number of frames used to derive the correlation coefficient,
The image matching apparatus according to claim 3, wherein: The derivation unit derives respective encoding difficulty levels when the first video signal is encoded at a plurality of encoding bit rates,
When the amount of time information for each frame of the first video signal is equal to or less than the first threshold, the classifying unit determines each encoding difficulty level when encoding at a plurality of encoding bit rates. Classifying the first video signal based on the characteristics;
When the amount of time information for each frame of the first video signal is equal to or less than the first threshold, the matching unit performs the mutual operation according to the classification result of the first video signal by the classification unit. Change the number of frames used to derive the correlation coefficient,
The image matching apparatus according to claim 3, wherein: Computer
A derivation procedure for deriving respective feature amounts of the first video signal and the second video signal in which the video quality of the first video signal has changed ;
A classification for classifying the first video signal into one of a plurality of types classified based on at least the magnitude of temporal change of the video based on the feature quantity of the first video signal. Procedure and
The first video signal and the second video signal based on the respective feature quantities of the first video signal and the second video signal corresponding to the classified type of the first video signal. An alignment procedure for calculating the amount of positional deviation with respect to
A video matching method comprising:   The program for functioning a computer as each part as described in any one of Claims 1 thru | or 6.

Images