JP5339855B2 - Motion vector search device, motion vector search method, image processing device, image processing method, program, and storage medium - Google Patents

Description

The present invention relates to a motion vector search device, a motion vector search method , an image processing device, an image processing method , a program, and a storage medium. The present invention relates to a technique suitable for use in a moving picture coding technique such as H.264.

  In recent years, digitalization of information related to multimedia such as audio signals and video signals has been rapidly progressing, and video signal compression encoding and decoding techniques have attracted attention accordingly. Since compression encoding and decoding techniques can reduce the storage capacity required for storing video signals and the bandwidth required for transmission, this technique is extremely important for the multimedia industry.

  These compression encoding and decoding techniques compress the amount of information / data amount by using the high autocorrelation (that is, redundancy) of many video signals. The redundancy of the video signal includes temporal redundancy and two-dimensional spatial redundancy, and temporal redundancy can be reduced by using block unit motion estimation and motion compensation. On the other hand, spatial redundancy can be reduced using discrete cosine transform (DCT).

  In MPEG systems known as compression encoding and decoding techniques, video frames / fields that change over time by using these techniques to reduce video signal redundancy (hereinafter, frames / fields are referred to as frames). The data compression effect is improved. Here, the above-described motion estimation in units of blocks for reducing temporal redundancy refers to the work of finding a block that is the closest approximation between a reference frame (past frame) continuously input and a frame to be encoded. It is. A vector representing the moving direction and moving amount of the corresponding block is called a motion vector. Therefore, motion estimation is synonymous with motion vector estimation.

  As a method for estimating a motion vector, a block matching method is generally used. The block matching method is a method in which two images of a reference frame and an encoding target frame are compared in units of blocks, and motion is estimated in units of blocks based on the degree of matching of signal types. With this block matching method, a motion vector is estimated for each block from the reference frame and the encoding target frame, and motion compensated prediction is performed using the estimated motion vector. This block matching method is described in Patent Document 1, for example.

JP-A-4-323780

  In the block matching method, generally, a vector having the smallest pixel difference between the block of the reference frame and the block of the encoding target frame is adopted as the motion vector. However, when the vector having the smallest absolute value of the pixel difference is a motion vector, the generated code amount of the macroblock can be reduced in many cases, but the encoding deterioration cannot be sufficiently dealt with.

  The reason why the encoding deterioration cannot be sufficiently dealt with will be described with reference to FIG. For example, in the MPEG system, the pixel difference between the current picture and the reference picture is orthogonally transformed for each block, quantized, and further decoded by inverse quantization and inverse orthogonal transform. In the description of FIG. 8, the orthogonal transform and the inverse orthogonal transform process are not related to the main subject, and thus the description thereof is omitted. Further, since the quantization and inverse quantization processes are not related to the main subject, detailed explanation is omitted, and only division and integration will be described. At that time, in FIG. 8, the amount of information is compressed by dividing the pixel difference value by 8 in quantization, and the information is restored by being multiplied by 8 in inverse quantization. In FIG. 8, description will be given focusing on one pixel in a block as a processing unit.

A case where the current pixel value 40a is 108 and the reference pixel value 41a is 100 will be described with reference to FIG.
In FIG. 8A, the difference value = 8 is obtained as a result of subtracting the reference pixel value 41a from the current pixel value 40a by the subtractor 42. Then, the result is transmitted to the quantization unit 43.

  In the quantization unit 43, the difference value = 8 is quantized (divided by 8), so that the quantized value = 1. Then, the result is transmitted to the inverse quantization unit 44. In the inverse quantization unit 44, the quantized value = 1 is inversely quantized (multiplied by 8), and as a result, the inverse quantized value = 8. Here, when the difference value and the value after inverse quantization are compared, since both values are 8, there is no amount of information lost by quantization, and the difference value = 8 can be completely restored by inverse quantization.

Next, a case where the current pixel value 40b is 108 and the reference pixel value 41b is 103 will be described with reference to FIG.
In FIG. 8B, the difference value = 5 is obtained as a result of subtracting the reference pixel value 41b from the current pixel value 40b by the subtractor 42. Then, the result is transmitted to the quantization unit 43.

  In the quantization unit 43, the difference value = 5 is quantized (divided by 8), so that the quantized value = 0. Then, the result is transmitted to the inverse quantization unit 44. In the inverse quantization unit 44, the quantized value = 0 is inversely quantized (multiplied by 8), so that the inverse quantized value = 0. Here, when the difference value is compared with the value after inverse quantization, the information amount cannot be completely restored by inverse quantization, and (difference value = 5) − (value after inverse quantization = 0) = 5 by quantization. Only the amount of information is lost.

  As described above, in FIG. 8A, the difference value = 8, whereas in FIG. 8B, the difference value = 5 and the difference value is small. However, in the case of FIG. 8A, the information can be completely restored, whereas in the case of FIG. 8B, the information cannot be completely restored and the image quality of the image is deteriorated. . As described above, there is a problem that encoding deterioration occurs due to the difference value.

  An object of the present invention is to perform encoding with a high compression rate and to suppress deterioration of image quality in view of the above-described problems.

The motion vector search apparatus of the present invention searches for a motion vector of a block by searching a plurality of blocks constituting a picture to be encoded in a search area set in a reference picture corresponding to the block. A motion vector search apparatus that calculates a difference between a block to be encoded and a pixel at a search position of the reference picture, and a difference between the block to be encoded and a pixel at a search position of a reference picture. It is characterized by having motion vector search means for searching for a motion vector of the block to be encoded based on the difference between the quantized and inverse quantized pixel and the sum of absolute values of the difference .

The motion vector search method of the present invention searches for a motion vector of a block by searching a plurality of blocks constituting a picture to be encoded in a search area set in a reference picture corresponding to the block. And a difference between a block to be encoded and a pixel at a search position of the reference picture, and a difference between the block to be encoded and a pixel at a search position of a reference picture. It has a motion vector search step of searching for a motion vector of the block to be encoded based on the difference between the quantized and inverse quantized pixel and the sum of absolute values of the difference .

  According to the present invention, it is possible to improve the encoding efficiency and suppress degradation of image quality.

(First embodiment)
Hereinafter, embodiments of the present invention will be described with reference to the drawings. First, in order to describe a conventional block matching method, the configuration of a conventional motion vector search device will be described.
FIG. 6 is a block diagram illustrating a configuration example of a motion vector search apparatus using a conventional block matching method.
In FIG. 6, the motion vector search apparatus includes a current frame / field storage unit 20 that stores a picture to be encoded, and a reference frame / field storage unit 21 that stores a reference picture. Furthermore, a current macroblock storage buffer 22 that stores a pixel block (macroblock) to be encoded among pictures to be encoded, a search window storage buffer 23, and a motion vector search unit 24 are provided.

  The current macroblock storage buffer 22 is used to extract an image of a macroblock to be encoded (current macroblock) from the current frame / field storage unit 20. The search window storage buffer 23 sets the center of the search area at the center of the current macroblock, and extracts and stores an image from the reference frame / field storage unit 21 for the range of the search area (hereinafter referred to as a search window). Is for.

  The motion vector search unit 24 takes the difference between the pixel value of the current macroblock and the pixel value of the reference frame / field (hereinafter referred to as a frame, representative of the frame / field) for each pixel in the search window. Then, a vector having the smallest absolute value sum of pixel differences is searched, and a final motion vector for the current macroblock is estimated.

  In addition, as shown in FIG. 7, the block matching method includes a configuration in which a motion vector storage unit 35 that stores the motion vector estimated by the motion vector search unit 24 is added to the motion vector search device of FIG. . In the configuration illustrated in FIG. 7, the motion vector estimated for the immediately preceding macroblock from the motion vector storage unit 35 is provided to the reference frame / field storage unit 21. The reference frame / field storage unit 21 sets the center position of the search region at a position shifted from the macroblock to be encoded by the motion vector of the previous macroblock, and stores the reference image in the search window storage buffer 23. Is output.

  The local motion compensation method described above is H.264, which is currently an international standard method. 261, H.H. It is used in H.263, MPEG1, MPEG2 and MPEG4 and is the most widely adopted motion compensation method.

  Hereinafter, a preferred embodiment of a motion vector search apparatus according to the present invention will be described in detail with reference to the block diagram of FIG. 1 and the flowchart of FIG. In the present embodiment, the operation of the motion vector search unit 14 is different from the conventional one.

FIG. 1 is a block diagram illustrating a configuration example of a motion vector search apparatus 50 according to the present embodiment.
In FIG. 1, a current frame / field storage unit 10 is for storing a picture to be encoded. The reference frame / field storage unit 11 stores a reference picture.

  The current macroblock storage buffer 12 is for extracting and storing an image of a macroblock to be encoded (current macroblock) from the current frame / field storage unit 20. The search window storage buffer 13 sets the center of the search area at the center of the current macroblock, and extracts and stores an image from the reference frame / field storage unit 21 for the range of the search area (search window). is there.

  The motion vector search unit 14 is for estimating a final motion vector within the search window. Hereinafter, a detailed configuration of each block configuring the motion vector search unit 14 will be described.

  The subtracter 100 is for obtaining a difference between the macroblock to be encoded and the reference frame at the search position in the search window. The orthogonal transform unit 101 is for performing integer orthogonal transform on the difference value obtained by the subtracter 100. The quantization unit 102 is for performing quantization at a predetermined quantization scale on the transform coefficient subjected to the integer orthogonal transform in the orthogonal transform unit 101.

  The inverse quantization unit 103 is for performing a predetermined inverse quantization process on the transform coefficient quantized by the quantization unit 102. The inverse orthogonal transform unit 104 performs inverse integer orthogonal transform that returns the transform coefficient inversely quantized by the inverse quantization unit 103 to the original image data space.

  The loss information amount calculation unit 105 obtains a difference value between the macroblock to be encoded restored by the inverse orthogonal transform in the inverse orthogonal transform unit 104 and the reference frame of the search position in the search window. The loss information amount calculation unit 105 also obtains a difference value between the encoding target macroblock and the reference frame at the search position in the search window, and further obtains a difference (loss information amount) between the two obtained difference values. .

  The evaluation function calculation unit 106 determines a motion vector using the difference value between the macro block to be encoded and the reference frame at the search position in the search window and the loss information amount obtained by the loss information amount calculation unit 105. This is for calculating the evaluation function cost for The motion vector determination unit 107 is a motion vector determination unit that determines a search position in a search window having a small evaluation function cost as a final motion vector. The difference absolute value sum calculation unit 108 is for calculating the absolute value sum of the difference values for each pixel obtained by the subtracter 100.

  Next, the flow of vector search processing in this embodiment will be described with reference to the block diagram of FIG. 1 and the flowchart of FIG. FIG. 2 is a flowchart showing an example of a motion vector search processing procedure in the present embodiment. In the present embodiment, in particular, the processing of the motion vector search unit 14 will be described in detail.

  First, in step S200, the subtracter 100 obtains a pixel-by-pixel difference between the encoding target macroblock and the reference frame at the search position in the search window. Then, the difference data is transmitted to the difference absolute value sum calculation unit 108, the loss information amount calculation unit 105, and the orthogonal transform unit 101. Next, in step S201, the orthogonal transform unit 101 performs integer orthogonal transform on the pixel-by-pixel difference obtained by the subtractor 100, and performs orthogonal transform from a general image space to a transform coefficient space with increased energy concentration. To the quantization unit 102.

  Next, in step S <b> 202, the quantization unit 102 performs quantization according to the orthogonal transform component with a predetermined step width on the transform coefficient of the orthogonally transformed difference data, and transmits the result to the inverse quantization unit 103. In step S <b> 204, the inverse quantization unit 103 restores the transform coefficient by performing inverse quantization on the quantized transform coefficient, and transmits the transform coefficient to the inverse orthogonal transform unit 104.

  Next, in step S204, the inverse orthogonal transform unit 104 returns the transform coefficient restored by the inverse quantization unit 103 to the original image data space, and restores the difference data from the predicted image information. Then, the restored difference data is transmitted to the loss information amount calculation unit 105. The restored difference data is an image that is slightly degraded from the input image data due to the quantization error in the quantization process.

  Next, in step S205, the loss information amount calculation unit 105 calculates the absolute value of the difference between the difference data for each pixel obtained in the subtracter 100 in step S200 and the difference data restored in step S204 as the coding deterioration degree. Calculates the value sum LOST. Then, the information of the absolute value sum LOST is transmitted to the evaluation function calculation unit 106. Note that the absolute value sum LOST obtained in step S205 represents the degree of coding deterioration obtained by quantifying the amount of information lost due to the quantization error in the quantization process.

  Next, in step S <b> 206, the difference absolute value sum calculation unit 108 calculates an absolute value sum MAD (Mean Absolute Difference) of the difference values for each pixel obtained by the subtracter 100, and transmits the result to the evaluation function calculation unit 106. Then, the evaluation function calculation unit 106 uses the absolute value sum MAD transmitted from the difference absolute value sum calculation unit 108 and the absolute value sum LOST transmitted from the loss information amount calculation unit 105 to calculate the square shown in the following equation (1). The evaluation function arithmetic expression of is calculated. Then, an evaluation function cost for determining a motion vector is calculated and transmitted to the motion vector determination unit 107.

  Here, α and β are real numbers for normalizing the dimensions of the absolute value sums MAD and LOST. In the expression shown in Equation 1, the motion vector that minimizes the distance when MAD = 0 and LOST = 0 and minimizes the distance when the absolute value sums MAD and LOST are taken on two axes is the optimal motion vector. Therefore, if the MAD is the same value, a motion vector search is performed so that the absolute value sum LOST is minimized.

  The evaluation function cost is calculated using the formula shown in the following formula 2, and the sum when the dimension of MAD = 0 and LOST = 0 is minimized and the dimension of the absolute value sum MAD and LOST is normalized is the smallest. This motion vector may be an optimal motion vector.

  Next, in step S207, the motion vector determination unit 107 determines the evaluation function cost transmitted from the evaluation function calculation unit 106 and the minimum evaluation function costmin that is the minimum value of the evaluation functions transmitted so far. And compare. As a result of the comparison, if the evaluation function cost is smaller, the process proceeds to step S208. In step S208, the minimum evaluation function costmin is updated to cost, and the final motion vector candidate vec is updated to the search position at that time. Note that the minimum evaluation functions costmin and vec are unconditionally updated at the first search position of the search window. On the other hand, as a result of the comparison in step S207, if the minimum evaluation function costmin is smaller or the same, the process proceeds to step S209.

  Next, in step S209, it is determined whether or not a motion vector has been searched for all the pixels in the search window. If the result of this determination is that all pixels have not been searched, the process returns to step S200 and the same processing is repeated. On the other hand, if it is determined in step S209 that motion vectors have been searched for all pixels, the process proceeds to step S210, where the vec at that time is determined as the final motion vector, and the process ends.

  As described above, according to the present embodiment, the evaluation function is related not only to the absolute value of the difference between the current macroblock and the reference image, which is a value correlated with the generated code amount, but also to the coding degradation. Contains a value. Therefore, it is possible to search for a motion vector that achieves both encoding efficiency and image quality.

  In this embodiment, the sum of absolute differences between the pixel of the encoding target frame and the decoded reference frame pixel is used as a value correlated with the encoding deterioration. On the other hand, the reciprocal of PSNR (peak signal-to-noise ratio) is calculated and used from the difference between the pixel of the encoding target frame and the decoded reference frame pixel, using the following equation (3). Also good.

  N and M are the numbers of pixels in the vertical and horizontal directions of the image, respectively, and p (i, j) is the pixel value of the location (i, j) of the current input image. Further, p ′ (i, j) is the pixel value of the location (i, j) of the restored image, and T is the number of gradations of the image−1 (T = 255 for an 8-bit / pixel image).

(Second Embodiment)
Hereinafter, the motion vector search apparatus according to the present embodiment will be described in detail with reference to FIG. The motion vector search apparatus according to the present embodiment is different from the first embodiment in that an orthogonal transform coefficient absolute value sum calculating unit 308 is provided instead of the difference absolute value sum calculating unit 108, but other configurations are provided. Is substantially the same as in the first embodiment. Since the operations other than the orthogonal transform coefficient absolute value sum calculation unit 308 and the evaluation function calculation unit 106 are the same as those in the first embodiment, the description thereof will be omitted.

  The orthogonal transform coefficient absolute value sum calculation unit 308 receives the orthogonal transform coefficient of the difference data from the orthogonal transform unit 101. Then, an absolute value sum SATD (Sum of Absolute Transform Difference) of the orthogonal transform coefficients is calculated and transmitted to the evaluation function calculation unit 106.

  Next, the evaluation function calculation unit 106 uses the absolute value sum SATD transmitted from the orthogonal transform coefficient absolute value sum calculation unit 608 and the absolute value sum LOST transmitted from the loss information amount calculation unit 105 to The square evaluation function arithmetic expression shown in 4 is calculated. Then, an evaluation function cost for determining a motion vector is calculated and transmitted to the motion vector determination unit 107.

Here, α and β are real numbers for normalizing the dimensions of the absolute value sums SATD and LOST.
In the equation shown in Equation 4, the motion vector that minimizes the distance when SATD = 0 and LOST = 0 is set to the minimum and the distance when the absolute value sum SATD and LOST are taken on two axes is the optimal motion vector.

  The evaluation function cost is calculated using the equation shown in the following formula 5, and the sum when the dimension of the absolute value sum SATD and LOST is normalized is minimized when SATD = 0 and LOST = 0. This motion vector may be an optimal motion vector.

  Next, the motion vector determination unit 107 compares the evaluation function cost transmitted from the evaluation function calculation unit 106 with the minimum evaluation function costmin that is the minimum value of the evaluation functions transmitted so far. As a result of this comparison, if the evaluation function cost is smaller, the minimum evaluation function costmin is updated to cost, and the final motion vector candidate vec is updated to the search position at that time. Note that the minimum evaluation functions costmin and vec are unconditionally updated at the first search position of the search window. Then, after performing the above processing for all the pixels in the search window, the vec at that time is determined as the final motion vector.

  As described above, according to the present embodiment, instead of the absolute value sum MAD, which is a value correlated with the generated code amount, the evaluation function uses the absolute value sum SATD, which has a stronger correlation with the generated code amount. I made it. As a result, it is possible to search for a motion vector that further considers encoding efficiency and achieves both image quality.

(Third embodiment)
Hereinafter, the motion vector search apparatus according to the present embodiment will be described in detail with reference to FIG. The motion vector search apparatus according to the present embodiment is different from the second embodiment in that it includes a binary code amount calculation unit 408 instead of the orthogonal transform coefficient absolute value sum calculation unit 308, but the other configurations are the same. This is almost the same as in the second embodiment. Since operations other than the binary code amount calculation unit 408 and the evaluation function calculation unit 106 are the same as those in the second embodiment, description thereof is omitted. In this embodiment, H.264 is used as the encoding method. 264 is assumed.

  The binary code amount calculation unit 708 receives the value coef obtained by quantizing the orthogonal transform coefficient from the quantization unit 102, calculates the binary code amount of the value coef, and transmits it to the evaluation function calculation unit 106. The evaluation function calculation unit 106 uses the binary code amount transmitted from the binary code amount calculation unit 708 and the absolute value sum LOST transmitted from the loss information amount calculation unit 105 to evaluate the square shown in the following Equation 6. Calculate a function expression. Then, an evaluation function cost for determining a motion vector is calculated and transmitted to the motion vector determination unit 107.

  Here, α and β are real numbers for normalizing the dimension of the binary code amount and the absolute value sum LOST. In the equation shown in Equation 6, the motion vector that minimizes the distance when the binary code amount = 0 and LOST = 0 is minimized and the distance when the binary code amount and the absolute value sum LOST are taken on two axes is the optimum. Let it be a motion vector.

  Note that the evaluation function cost is calculated using the equation shown in Equation 7 below, and when the binary code amount = 0 and LOST = 0 is minimized, the dimensions of the binary code amount and the absolute value sum LOST are normalized. The motion vector with the smallest sum of time may be set as the optimum motion vector.

  Next, the motion vector determination unit 107 compares the evaluation function cost transmitted from the evaluation function calculation unit 106 with the minimum evaluation function costmin that is the minimum value of the evaluation functions transmitted so far. As a result of this comparison, if the evaluation function cost is smaller, the minimum evaluation function costmin is updated to cost, and the final motion vector candidate vec is updated to the search position at that time. Note that the minimum evaluation functions costmin and vec are unconditionally updated at the first search position of the search window. Then, after performing the above processing for all the pixels in the search window, the vec at that time is determined as the final motion vector.

  As described above, according to the present embodiment, the evaluation function uses a binary code amount having a strong correlation with the generated code amount instead of the absolute value sum SATD that is a value correlated with the generated code amount. I made it. As a result, it is possible to search for a motion vector that further considers encoding efficiency and achieves both image quality.

(Fourth embodiment)
Hereinafter, the motion vector search apparatus according to the present embodiment will be described in detail with reference to FIG. The motion vector search apparatus according to the present embodiment is different from the third embodiment in that it includes a vector code amount calculation unit 809 and a motion vector storage unit 810, but other configurations are substantially the same as those of the second embodiment. The same. Since operations other than the vector code amount calculation unit 809, the motion vector storage unit 810, and the evaluation function calculation unit 106 are the same as those in the third embodiment, description thereof is omitted.

  The vector code amount calculation unit 809 reads the reference frame at the search position in the search window from the search window storage buffer 13 and receives the motion vectors of the peripheral macroblocks obtained in advance from the motion vector storage unit 810. Then, the code amount VEC_CODE of the motion vector is calculated and transmitted to the evaluation function calculation unit 106.

  The evaluation function calculation unit 106 uses the binary code amount transmitted from the binary code amount calculation unit 708, the vector code amount VEC_CODE described above, and the absolute value sum LOST transmitted from the loss information amount calculation unit 105 as follows. A square evaluation function arithmetic expression shown in Equation 8 is calculated. Then, an evaluation function cost for determining a motion vector is calculated and transmitted to the motion vector determination unit 107.

  Here, α, β, and γ are real numbers for normalizing the dimensions of the binary code amount, the absolute value sum LOST, and the vector code amount VEC_CODE. The equation shown in Equation 8 is obtained when the binary code amount = 0, LOST = 0, and VEC_CODE = 0 are minimized, and the binary code amount, the absolute value sum LOST, and the vector code amount VEC_CODE are taken on three axes. The motion vector that makes the distance the shortest is the optimal motion vector.

  Since both the binary code amount and the vector code amount VEC_CODE are correlated values as the code amount, the evaluation function cost may be calculated using the following equation (9) by combining the two values. .

  Note that the evaluation function cost may be calculated using the following equation (10). In this case, the sum when the binary code amount = 0, LOST = 0, VEC_CODE = 0 is minimized, and the dimensions of the binary code amount, the absolute value sum LOST, and the vector code amount VEC_CODE are normalized is the smallest. This motion vector may be an optimal motion vector.

  Furthermore, since the binary code amount and the vector code amount VEC_CODE are both correlated values as the code amount, the two values may be combined and the evaluation function cost may be calculated using the following equation (11). .

  The motion vector determination unit 107 compares the evaluation function cost transmitted from the evaluation function calculation unit 106 with the minimum evaluation function costmin that is the minimum value of the evaluation functions transmitted so far. As a result of this comparison, if the evaluation function cost is smaller, the minimum evaluation function costmin is updated to cost, and the final motion vector candidate vec is updated to the search position at that time. Note that the minimum evaluation functions costmin and vec are unconditionally updated at the first search position of the search window. Then, after performing the above processing for all the pixels in the search window, the vec at that time is determined as the final motion vector and stored in the motion vector storage unit 810.

  As described above, according to the present embodiment, by enhancing the correlation with the code amount generated by adding the vector code amount to the evaluation function, it is possible to obtain a motion vector that further considers encoding efficiency and achieves both image quality. It becomes possible to search.

(Other embodiments according to the present invention)
Each means constituting the motion vector search apparatus and each step of the motion vector search method in the embodiment of the present invention described above can be realized by operating a program stored in a RAM or ROM of a computer. This program and a computer-readable storage medium storing the program are included in the present invention.

  In addition, the present invention can be implemented as, for example, a system, apparatus, method, program, storage medium, or the like. Specifically, the present invention may be applied to a system including a plurality of devices. The present invention may be applied to an apparatus composed of a single device.

  The present invention includes a case where a software program (in the embodiment, a program corresponding to the flowchart shown in FIG. 2) for realizing the functions of the above-described embodiments is directly or remotely supplied to a system or apparatus. This includes the case where the system or the computer of the apparatus is also achieved by reading and executing the supplied program code.

  Accordingly, since the functions of the present invention are implemented by computer, the program code installed in the computer also implements the present invention. In other words, the present invention includes a computer program itself for realizing the functional processing of the present invention.

  In that case, as long as it has the function of a program, it may be in the form of object code, a program executed by an interpreter, script data supplied to the OS, and the like.

  Examples of the storage medium for supplying the program include a flexible disk, a hard disk, an optical disk, and a magneto-optical disk. Further, there are MO, CD-ROM, CD-R, CD-RW, magnetic tape, nonvolatile memory card, ROM, DVD (DVD-ROM, DVD-R) and the like.

  As another program supply method, there is a method of connecting to a homepage on the Internet using a browser of a client computer. It can also be supplied by downloading the computer program itself of the present invention or a compressed file including an automatic installation function from a homepage to a storage medium such as a hard disk.

  It can also be realized by dividing the program code constituting the program of the present invention into a plurality of files and downloading each file from a different homepage. That is, a WWW server that allows a plurality of users to download a program file for realizing the functional processing of the present invention on a computer is also included in the present invention.

  As another method, the program of the present invention is encrypted, stored in a storage medium such as a CD-ROM, distributed to users, and encrypted from a homepage via the Internet to users who have cleared predetermined conditions. Download the key information to be solved. It is also possible to execute the encrypted program by using the key information and install the program on a computer.

  Further, the functions of the above-described embodiments are realized by the computer executing the read program. Furthermore, based on the instructions of the program, an OS or the like running on the computer performs part or all of the actual processing, and the functions of the above-described embodiments can be realized by the processing.

  As another method, a program read from a storage medium is first written in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer. Then, based on the instructions of the program, the CPU or the like provided in the function expansion board or function expansion unit performs part or all of the actual processing, and the functions of the above-described embodiments are also realized by the processing.

It is a block diagram which shows the structural example of the motion vector search apparatus which concerns on the 1st Embodiment of this invention. It is a flowchart which shows an example of the search process procedure of the motion vector in the 1st Embodiment of this invention. It is a block diagram which shows the structural example of the motion vector search apparatus which concerns on the 2nd Embodiment of this invention. It is a block diagram which shows the structural example of the motion vector search apparatus which concerns on the 3rd Embodiment of this invention. It is a block diagram which shows the structural example of the motion vector search apparatus which concerns on the 4th Embodiment of this invention. It is a block diagram which shows the structural example of the motion vector search apparatus using the conventional block matching system. It is a block diagram which shows the other structural example of the motion vector search apparatus using the conventional block matching system. It is a figure explaining performing the orthogonal transformation etc. of the pixel difference of a current picture and a reference picture for every block, and a coding degradation arises.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Current frame / field storage unit 11 Reference frame / field storage unit 12 Current macroblock storage buffer 13 Search window storage buffer 14 Motion vector search unit 50 Motion vector search device 100 Subtractor 101 Orthogonal transform unit 102 Quantization unit 103 Inverse quantization Unit 104 inverse orthogonal transform unit 105 loss information amount computation unit 106 evaluation function computation unit 107 motion vector determination unit 108 difference absolute value sum computation unit 308 orthogonal transform coefficient absolute value sum computation unit 408 binary code amount computation unit 509 vector code amount computation Unit 510 motion vector storage unit

Claims (17)

A motion vector search device for searching for a motion vector of a block by searching for a plurality of blocks constituting a picture to be encoded in a search region set in a reference picture corresponding to the block,
The difference between the difference between the block to be encoded and the pixel at the search position of the reference picture and the difference between the block to be encoded and the pixel at the search position of the reference picture are quantized and dequantized. A motion vector search apparatus comprising: a motion vector search means for searching for a motion vector of a block to be encoded based on a difference and a sum of absolute values of the differences . The motion vector search means is the encoding target based on the absolute value sum of the difference and the absolute value sum of the difference between the block to be encoded and the pixel at the search position of the reference picture. The motion vector search apparatus according to claim 1 , wherein a motion vector of a block is searched. The motion vector search means performs the encoding based on the absolute value sum of the differences and an orthogonal transform coefficient obtained by orthogonally transforming the difference between the block to be encoded and the pixel at the search position of the reference picture. The motion vector search apparatus according to claim 1 , wherein a motion vector of a target block is searched. The motion vector search means is based on a sum of absolute values of the differences and a binary code amount of a value obtained by quantizing a difference between a block to be encoded and a pixel at a search position of the reference picture. The motion vector search apparatus according to claim 1 , wherein a motion vector of a block to be encoded is searched. The motion vector search unit further claim 4, characterized in that searching for a motion vector of a block to be the encoding based on the code amount of motion vectors of a block to be the encoding Motion vector search device. A motion vector search method for searching a motion vector of a block by searching for a plurality of blocks constituting a picture to be encoded in a search region set in a reference picture corresponding to the block,
The difference between the difference between the block to be encoded and the pixel at the search position of the reference picture and the difference between the block to be encoded and the pixel at the search position of the reference picture are quantized and dequantized. A motion vector search method comprising a motion vector search step of searching for a motion vector of a block to be encoded based on a difference and a sum of absolute values of the differences . A program that causes a computer to execute a search for a motion vector of a block by searching for a plurality of blocks constituting a picture to be encoded in a search region set in a reference picture corresponding to the block. There,
The difference between the difference between the block to be encoded and the pixel at the search position of the reference picture and the difference between the block to be encoded and the pixel at the search position of the reference picture are quantized and dequantized. A program for causing a computer to execute a motion vector search step for searching for a motion vector of a block to be encoded based on a difference and a sum of absolute values of differences . A computer-readable storage medium storing the program according to claim 7 . An image processing apparatus that determines a motion vector corresponding to an image of a block to be encoded using an image of the block to be encoded and an image of a reference picture corresponding to the block to be encoded,
For each of the plurality of positions of the reference picture image,
Calculating first difference data that is a difference between an image of the block to be encoded and an image at each position of the reference picture;
Quantizing the first difference data;
Dequantizing the quantized first difference data;
The encoding is performed based on a difference value between the dequantized first difference data and the unquantized first difference data corresponding to each of a plurality of positions of the reference picture image. An image processing apparatus for determining a motion vector corresponding to a target block. Based on the sum of absolute values of the difference values between the dequantized first difference data and the unquantized first difference data corresponding to each of a plurality of positions of the reference picture image, The image processing apparatus according to claim 9 , wherein a motion vector corresponding to a block to be encoded is determined. The absolute value sum of the difference values between the dequantized first difference data and the unquantized first difference data corresponding to each of a plurality of positions of the reference picture image, and the encoding target The image processing according to claim 10 , wherein a motion vector corresponding to the block to be encoded is determined based on an absolute value sum of difference values between the image of the block and the image of the reference picture. apparatus. The absolute value sum of the difference values between the dequantized first difference data and the unquantized first difference data corresponding to each of a plurality of positions of the reference picture image, and the encoding target A motion vector corresponding to the block to be encoded is determined based on an absolute value sum of orthogonal transform coefficients obtained by orthogonally transforming a difference value between the image of the block and the reference picture. Item 15. The image processing apparatus according to Item 10 . The absolute value sum of the difference values between the dequantized first difference data and the unquantized first difference data corresponding to each of a plurality of positions of the reference picture image, and the encoding target The motion vector corresponding to the block to be encoded is determined based on a binary code amount obtained by quantizing a difference value between the image of the block and the image of the reference picture. The image processing apparatus according to 10 . The absolute value sum of the difference values between the dequantized first difference data and the unquantized first difference data corresponding to each of a plurality of positions of the reference picture image, and the encoding target The encoding target based on a binary code amount obtained by quantizing a difference value between the image of the block and the reference picture and a code amount of a motion vector corresponding to each position of the reference picture image The image processing apparatus according to claim 13 , wherein a motion vector corresponding to each block is determined. An image processing method for determining a motion vector corresponding to an image of a block to be encoded using an image of the block to be encoded and an image of a reference picture corresponding to the block to be encoded,
For each of the plurality of positions of the reference picture image,
Calculating first difference data that is a difference between an image of the block to be encoded and an image at each position of the reference picture;
Quantizing the first difference data;
Dequantizing the quantized first difference data;
The encoding is performed based on a difference value between the dequantized first difference data and the unquantized first difference data corresponding to each of a plurality of positions of the reference picture image. An image processing method characterized by determining a motion vector corresponding to a target block. Based on the sum of absolute values of the difference values between the dequantized first difference data and the unquantized first difference data corresponding to each of a plurality of positions of the reference picture image, The image processing method according to claim 15 , wherein a motion vector corresponding to a block to be encoded is determined. A program for operating a computer to determine a motion vector corresponding to an image of a block to be encoded using an image of the block to be encoded and an image of a reference picture corresponding to the block to be encoded There,
For each of the plurality of positions of the reference picture image,
Calculating first difference data that is a difference between an image of the block to be encoded and an image at each position of the reference picture;
Quantizing the first difference data;
Dequantizing the quantized first difference data;
The encoding is performed based on a difference value between the dequantized first difference data and the unquantized first difference data corresponding to each of a plurality of positions of the reference picture image. A program that causes a computer to execute a step of determining a motion vector corresponding to a target block.

Images