JP2007208432A - Moving picture decoder and decoding program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide moving picture decoder and decoding program which can produce a more naturally decoded image by processing the errors generated, when an encoded moving picture is decoded more efficiently. <P>SOLUTION: Upon the occurrence of errors in a non-reference picture, a correlation judging section 113 judges existence of correlation, by comparing the occupation ratio of intra MB in the picture with a threshold F, respectively, for the reference pictures immediately before and after that error picture. Furthermore, average value and variance value of the MV value of each MB of these reference pictures are calculated, and a reference picture of smaller average value and variance value is selected as a picture, having correlation the most. Since the MV value can be complemented with the MV value of a reference picture having correlation, even if errors occur in the non-reference picture immediately before occurrence of scene change, the decoded image can move more naturally. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a moving picture decoding apparatus and a moving picture decoding program, and particularly occurs when decoding moving picture encoded data created using block-unit orthogonal transform, quantization, and variable length coding. The present invention relates to a moving picture decoding apparatus and a moving picture decoding program that perform error processing.

  The encoded video data generated using block-level orthogonal transform, quantization, and variable length coding is decoded, dequantized, and inverse orthogonal transformed for the digital video signal to be encoded. When a block error is detected in a moving picture decoding apparatus using the decoding apparatus to decode a moving picture signal, if it is determined that the block and the next neighboring block to be decoded cannot be decoded independently, the neighboring block For example, there is known a video decoding device that sets a motion vector of a block immediately above the screen of the block in which the above error is detected and a motion vector of the block in which the above error is detected (for example, Patent Literature 1).

  FIG. 5 shows a block diagram of an example of a conventional moving picture decoding apparatus described in Patent Document 1. In the figure, a decoding unit 501 decodes input encoded data. Here, in the encoded data, one frame is divided for each block, encoded for each block, and encoded by a method in which motion compensation prediction is performed for each block between temporally neighboring frames. . The error detection unit 502 detects whether there is an error during decoding by the decoding unit 501. When no error is detected, the motion vector of the block currently decoded by the decoding unit 501 is stored in the motion vector storage unit 505 and output to the motion compensation decoding unit 507 via the motion vector setting unit 506 as it is.

  The motion compensation decoding unit 507 reads predicted image data that is temporally nearby from the frame memory 508 according to the input motion vector, and outputs the prediction image data to the adder 504. The adder 504 adds the decoded difference image data from the decoding unit 501 and the predicted image data from the motion compensation decoding unit 507, generates decoded image data, and outputs the decoded image data to the frame memory 508. Store.

  When an error is detected by the error detection unit 502, in order to suppress the influence of the error, the error determination unit 503 analyzes the characteristics of the blocks after the block in which the error is detected. If the block can be decoded, an error determination signal indicating that normal decoding can be performed is sent to the decoding unit 501 because blocks subsequent to the block in which the error is detected can be normally decoded. Process only the current error block.

  On the other hand, if the decoding cannot be performed independently of the current error block as a result of the analysis, an error determination signal indicating the fact is supplied to the decoding unit 501, and the decoding unit 501 has the next decoding unit that is not affected by the current error. All the blocks up to immediately before are regarded as error blocks, and the motion vector (Motion vector) is transferred to the current error block (here, macroblock: MB) via the motion vector storage unit 505 and the motion vector setting unit 506. : MV) value.

  Next, a method for complementing the MV value related to the error MB will be described. The motion vector setting unit 506 determines whether or not an error MB has occurred at the picture boundary (step 701 in FIG. 7). If the error MB is not at the picture boundary, as shown in FIG. The MV value of the upper MB is used for complementation (step 702 in FIG. 7).

  However, if the error MB is at the top of the picture, there is no MV data to complement. Therefore, in Patent Document 1, when the erroneous MB is the top part of a picture, or in the case of a picture unit, that is, at a picture boundary, (0, 0) is used as the MV value to be complemented by the erroneous MB. This is set (step 703 in FIG. 7).

JP 2003-179937 A

  However, in the conventional moving picture decoding apparatus described in Patent Document 1, if an error in units of pictures (hereinafter also referred to as error pictures) occurs during decoding, or an error occurs in an MB at a slice boundary (hereinafter referred to as error). Since this is a technique for complementing (0, 0) as the MV value when the boundary MB occurs, the complemented image data is a previously decoded reference picture. In the case of a moving image, the same image Will continue several times, and there is a problem that the decoded image moves unnaturally.

  In addition, since the conventional moving picture decoding apparatus described in Patent Document 1 performs error processing without distinguishing between reference pictures and non-reference pictures, if an error occurs in a reference picture, the previous decoded picture is used as it is. There is also a problem that the influence on the non-reference picture and the reference picture later in the decoding order is large.

  The present invention has been made in view of the above points, and is a moving picture decoding apparatus capable of performing error processing generated when decoding a coded moving picture more efficiently and obtaining a more natural decoded picture. It is another object of the present invention to provide a moving picture decoding program.

In order to achieve the above object, the first invention divides each screen of a moving image signal into block units, which are image regions having a predetermined number of pixels, and performs orthogonal transform, quantization, and variable length coding on a block basis. Non-reference to be decoded with reference to the motion vector information of other pictures, and is composed of a plurality of blocks of picture data. Decoding, inverse quantization, and inverse orthogonal transformation are performed on moving picture encoded data including a picture or a reference picture whose motion vector information is referred to when decoding another picture. A moving image decoding device for generating image data,
Decoding means for decoding moving image encoded data and outputting decoded coefficient data and decoded MV data, and detecting whether or not an error has occurred during decoding of the moving image encoded data decoded by the decoding means An error detection unit; a decoding order control unit that controls a decoding order of moving image encoded data according to a picture type of the error in which the error is detected; By controlling the conversion order, in accordance with a control signal output from the decoding order control means, MV value storage means for holding the decoded MV data output from the decoding means, and at least MV data from the MV value storage means An MV calculation unit that calculates an MV value using the decoded MV data from the decoding unit, and a picture in which an error is detected by the error detection unit is a non-reference picture In the decoding order of the non-reference picture, one or more reference pictures on at least one of the front and rear sides are set as the complement target picture, and the intra block existing ratio in the complement target picture is predetermined. Correlation determining means for determining the presence or absence of correlation by comparing with a threshold value, and when the complement target picture determined to be correlated by the correlation determination means is not one, each of the plurality of complement target pictures At least one of the average value and the variance value of the MV values from the MV calculating means is calculated, and the complementary target picture having the smallest calculated value is determined as a picture having the complementary MV value, and there is a correlation by the correlation determining means. When there is one complementation target picture determined to be, a complementing picture determination that determines the complementation target picture as a picture having a complementation MV value If the image data obtained by performing inverse quantization and inverse transformation on the decoded coefficient data output from the stage and the decoding means is a non-reference picture, the decoded reference picture data and the MV value are calculated. Decoded image data generating means for adding the motion compensated image data generated from the MV value from the means to generate decoded image data and storing the decoded image data of the reference picture as decoded reference picture data; When generating the decoded image data of the picture in which the error is detected by the error detecting means, the decoded MV data of the picture having the complementary MV value determined by the complementary picture determining means is calculated by the image data generating means. And MV data complementing means for complementing as an MV value from the means.

  In order to achieve the above object, a moving picture decoding program according to a second invention causes a computer to function as each component of the moving picture decoding apparatus according to the first invention.

  In the moving picture decoding apparatus and moving picture decoding program of the present invention, when it is detected that an error has occurred in a non-reference picture to be decoded, the preceding and / or subsequent ones in the decoding order of the non-reference picture are detected. Alternatively, two or more reference pictures are used as complementing target pictures, and the existence ratio of intra blocks in the complementing target pictures is compared with a predetermined threshold to determine whether or not there is a correlation. When the complement target picture is not one, at least one of the average value and the variance value of the MV values from the MV calculation unit is calculated for each of the plurality of complement target pictures, and the complement target picture having the smallest calculated value is calculated. When a picture having a complementary MV value is determined as a picture to be complemented and determined to have correlation, the picture to be complemented is a pixel having a complementary MV value. Since the MV value of the picture having the determined complementary MV value is complemented as the MV value of the non-reference picture in which the error is detected, the scene change is performed before and after the non-reference picture in which the error is detected. Even when it occurs or when the scene changes greatly, an appropriate MV value can be selected and determined.

  According to the present invention, one or two or more reference pictures before and / or after in the decoding order of a non-reference picture in which an error is detected at the time of decoding are set as a complement target picture, and an intra block occupies the complement target picture By comparing the ratio with a predetermined threshold value to determine the presence or absence of correlation, and complementing the MV value of the correlated reference picture as the MV value of the non-reference picture in which the error is detected, Even when a scene change occurs before or after a non-reference picture in which an error is detected, or even when the scene change is large, an appropriate MV value can be selected and determined, so that a decoded image that moves more naturally can be obtained. .

  Next, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of an embodiment of a moving picture decoding apparatus according to the present invention, and FIG. 2 is a flowchart of an embodiment when a computer performs a decoding operation using the moving picture decoding program according to the present invention. . As shown in FIG. 1, in this embodiment, MPEG or the like using orthogonal transform, quantization, and variable length coding, which is an image area consisting of a plurality of pixels, on one screen of a moving image signal. A device that decodes a bit stream 101 that is a data sequence of moving image encoded data obtained by encoding with the above encoding method, and includes a memory 102 that holds the bit stream 101, and a moving image from the memory 102 A decoding unit 103 that decodes the image encoded data, an inverse quantization unit 104 that inversely quantizes the decoded data from the decoding unit 103, an inverse transformation unit 105 that performs inverse transformation on the orthogonal transformation on the coding unit side, A decoding order control unit 112 that instructs at least encoded data to be output to the memory 102 is included. In addition, the video decoding device includes an adder 107, an error processing unit 108, an MV calculation unit 109, an MC unit 110, a frame memory 111, a correlation determination unit 112, and an MV value storage unit 114. .

  The memory 102 outputs the encoded video data constituting the bit stream 101 to the decoding unit 103 based on the control signal from the decoding order control unit 112 for the held bit stream 101. The decoding unit 103 decodes the moving image encoded data input from the memory 102. If the decoding unit 103 cannot decode the moving image encoded data, that is, if an error occurs in the moving image encoded data in units of pictures or in arbitrary blocks in the uppermost position of the slice, the error processing unit 108. When the error is reported and decoding is performed, the coefficient data obtained by decoding is output to the inverse quantization unit 104, and the motion vector (MV) value data obtained by decoding is output to the MV calculation unit 109. Is output. The inverse quantization unit 104 inversely quantizes the coefficient data input from the decoding unit 103 and outputs the obtained data to the inverse transform unit 105 to perform an inverse transform from the orthogonal transform at the time of encoding.

  When the decoding unit 103 determines an error, the error processing unit 108 receives error picture data or error boundary MB data, and performs error processing to be described later. The MV calculation unit 109 uses the decoded MV value data and the MV value data input from the MV value storage unit 114, and calculates the MV value using the peripheral information of the calculated MB. In the non-reference picture, the MV value can be calculated using the MV values of the reference pictures before and after the non-reference picture. The MV calculation unit 109 outputs the calculated MV value to the MC unit 110. If the calculated MV value picture is a reference picture, the MV value data is also output to the MV value storage unit 114. The MV value complement in error processing will be described in detail later.

  The MC unit 110 creates an MC image (motion compensation image) using the MV value calculated by the MV calculation unit 109 and the reference picture input from the frame memory 111, and adds the created MC image to the adder 107. Is added to the data output from the inverse transform unit 105 and the decoded image output 106 is generated. The frame memory 111 holds the decoded image data input from the adder 107 and outputs the decoded image data to the MC unit 110 as necessary.

  Note that a reference picture is a picture whose motion vector information of the picture itself is referred to when decoding another picture, and is image data for one screen used for generating an MC image. The picture is a picture to be decoded with reference to motion vector information of another picture, and is image data for one screen that is not used to generate an MC image.

  The decoding order control unit 112 controls the output of moving image encoded data from the memory 102 to the decoding unit 103 when the error processing unit 108 needs to control the decoding order. The correlation determination unit 113 uses the data of the data structure 2 (described later) input from the error processing unit 108 and the MV value data input from the MV value storage unit 114 to correlate the error picture with the picture that complements the error. The data of the updated data structure 2 is created and output to the error processing unit 108. The operation of the correlation determining unit 113 will be described in detail later. The MV value storage unit 114 holds the MV value data input from the MV calculation unit 109 and outputs the held MV value data to the MV calculation unit 109 as necessary.

  In the present embodiment, the MV value data stored by the MV value storage unit 114 is the previous reference picture, but the present invention is not limited to this. In addition, (0, 0) is set as the MV value of the block for which in-plane prediction is performed. The data structure of the MV value data to be stored is data structure 1. Here, the data structure 1 and the data structure 2 are the following structures.

[Data structure 1]: (Intra / inter slice discrimination flag, picture number, picture
[Intra / inter, MV value (X, Y)]) for any number of blocks
[Data structure 2]: (error picture number, complementary MV picture number)
Next, the operation of the embodiment of FIG. 1 will be described with reference to the flowchart of FIG. The memory 102 in FIG. 1 stores an input bitstream 101 in which moving image encoded data is synthesized in time series (processing 1 in step 201 in FIG. 2), and the encoded moving image encoded data of the bitstream 101 is stored. Is output (processing 2 in step 202 in FIG. 2). Subsequently, the moving image encoded data input by the decoding unit 103 in FIG. 1 is decoded (processing 3 in step 203 in FIG. 2).

  The decoding unit 103 determines whether or not the above decoding has been performed normally (branch 1 in step 204 in FIG. 2), and determines that the decoding has not been performed normally (that is, determines that an error has occurred) ) Causes the error processing unit 108 to perform error processing, which will be described later, to create data of the data structure 2, and outputs the data to the correlation determination unit 113 (step 205 in FIG. 2).

  At this time, the decoding order control unit 112 is notified from the error processing unit 108 that the data being decoded is an error, and the picture of the decoded data in which the error has been detected is the MV value data of the picture itself. Is a reference picture to be referred to when decoding another picture (that is, an intra prediction encoded image (I picture) or a one-sided predictive encoded image (P picture) defined in MPEG)) At this time, the decoding control is performed to instruct the memory 102 to output the next decoded data to the decoding unit 103 in the decoding order of the error picture.

  The correlation determination unit 113 uses the data of the data structure 2 from the error processing unit 108 and the data of the data structure 1 input from the MV value storage unit 114 to detect the correlation of the image, and can be more effective. Data of data structure 2 that can be greatly complemented is created (correlation determination process in step 206 in FIG. 2). The correlation determination process of the correlation determination unit 113 will be described later in detail.

  Subsequently, the MV calculation unit 109 calculates the data structure 2 data calculated in the error process in step 205 or the correlation determination process in step 206 and the data structure 1 data stored in the MV value storage unit 114. The MV value to be complemented is calculated from the picture interval between the error picture and the picture to complement the error (Step 4 in Step 207 in FIG. 2).

  Subsequently, the MC unit 110 creates a motion compensated image (MC image) from the MV value from the MV calculation unit 109 calculated in step 207 and the image data of the reference picture read out from the frame memory 111. It outputs to the adder 107 and adds with the data from the inverse transformation part 105, and produces | generates a decoded image (process 5 of step 208 of FIG. 2).

  On the other hand, when it is determined in step 204 that decoding has been performed normally, general decoding processing represented by MPEG (Moving Picture Experts Group) is performed (processing 6 in step 209 in FIG. 2). That is, in process 6 of step 209, the coefficient data decoded and output from the decoding unit 103 is inversely quantized by the inverse quantization unit 104 and then inversely converted by the inverse conversion unit 105, and then is added to the adder 107. On the other hand, the MV data decoded and output from the decoding unit 103 is calculated as an MV value by the MV calculation unit 109 and then supplied to the MC unit 110, where a reference picture input from the frame memory 111 is used. Then, an MC image is created and supplied to the adder 107. As a result, the decoded image output 106 is output from the adder 107 and stored in the frame memory 111 as necessary.

  When the process 5 in step 208 or the process 6 in step 209 is completed, it is determined whether or not the created decoded image is a reference picture (branch 2 in step 210 in FIG. 2), and the created decoded image is a reference picture. Is determined, the MV value used when the MC image is generated in the MC unit 110 is stored in the MV value storage unit 114. At this time, (0, 0) is set as the MV value of the block for which in-plane prediction is performed. Set (processing 7 of step 211 in FIG. 2). If it is determined that the decoded image created in branch 2 of step 210 is not a reference picture (non-reference picture), the decoding process is terminated without performing the process 7. Note that a non-reference picture is a picture that is to be decoded with reference to motion vector information of another picture. In the case of MPEG, a non-reference picture is a bi-predictive encoded image (B picture).

  Next, the correlation determination unit 113 for calculating the correlation, which constitutes the main part of the present embodiment, will be described in detail with reference to FIG. The correlation determination unit 113 analyzes the data of the data structure 2 input from the error processing unit 108, and determines whether or not there is one complement target picture (branch 1 in step 301 in FIG. 3). Here, when the picture of the decoded data in which error is detected (error picture) is a reference picture, the complement target picture is when the immediately preceding reference picture has an MV value in the decoding order of the error picture. When the immediately preceding reference picture has no MV value (for example, in the case of an I picture), the reference picture has the immediately following MV value. Further, when the error picture is a non-reference picture and the reference pictures having MV values before and after the error picture are used at the time of decoding, the reference pictures before and after the reference pictures are the complement target pictures.

  When the correlation determination unit 113 determines in step 301 in FIG. 3 that there is one complement target picture, the correlation target unit 113 determines the complement target picture as the picture in step 4 of step 306 described later. On the other hand, if the correlation determination unit 113 determines that there are a plurality of complement target pictures in step 301, the intra MB existing ratio (hereinafter referred to as an intra rate) in each complement target picture is determined for each complement target picture. Calculate (Process 1 in Step 302 of FIG. 3).

  Subsequently, the correlation determining unit 113 compares the intra rate of each picture calculated in step 302 with a predetermined threshold value F, and if the relationship of intra rate <threshold value F is satisfied, the correlation rate is determined. When the relationship of intra rate ≧ threshold value F is satisfied, it is determined that there is no correlation (process 2 in step 303 in FIG. 3). Thereby, for example, if the complement target picture whose correlation should be determined is a picture after a scene change point or a picture that is difficult to predict with respect to an error picture, it is determined that there is no correlation, and a picture before the scene change point In the case of a picture that is easy to predict, it is determined that there is a correlation.

  Subsequently, the correlation determining unit 113 determines whether or not there is one complement target picture determined to have a correlation in step 303 (branch 2 in step 304 in FIG. 3). When there is one complementation target picture determined to have correlation, it is determined as a picture having an MV value that complements the complementation picture (process 4 in step 306 in FIG. 3).

  On the other hand, when the correlation determining unit 113 determines that there is not one complement target picture determined to be correlated in step 304, for each complement target picture, the MV values of a plurality of blocks constituting the picture Are calculated (step 3 in step 305 in FIG. 3), and a picture having an MV value that complements the picture having the smallest average value and variance value is determined (step 306 in FIG. 3). Process 4).

  However, if the picture with the smallest average value and the picture with the smallest variance value are different from the complement target pictures determined to have a plurality of correlations, the picture with the smaller average value of the MV values is The picture having the MV value to be complemented is determined. An example of this case is shown in Table 1.

表１に示すように、ピクチャＡ及びピクチャＢがともに相関性あると判断された補完対象ピクチャであると仮定する。算出した平均値がパターン１、つまりピクチャＡ＜ピクチャＢの場合、最も相関性があるピクチャとして決定されるのはピクチャＡとなる。また、パターン２は、ピクチャＡ＞ピクチャＢの場合であり、最も相関性があるピクチャとして決定されるのはピクチャＢとなる。

As shown in Table 1, it is assumed that the picture A and the picture B are complementary pictures that are determined to be correlated. When the calculated average value is pattern 1, that is, picture A <picture B, it is picture A that is determined as the most correlated picture. Pattern 2 is a case where picture A> picture B, and picture B is determined as the most correlated picture.

  Then, the data having the data structure 2 is created using the picture having the MV value to be complemented in step 306 and the error complement target picture determined (step 5 in step 307 in FIG. 3). Note that the method of determining a picture having a MV value to be complemented in step 306 in FIG. 3 is not limited to the method of selecting a complement target picture having a smaller average value, and a picture having a smaller variance value is selected. It may be possible to select one image from a plurality of complement target pictures determined to have correlation, but a plurality of images may be selected, and an average value of the MV values, etc. It is also possible to complement the MV value of the error picture.

  Next, an example in which the operation of the moving picture decoding apparatus of the present invention shown in FIG. 1 is compared with the operation of a conventional moving picture decoding apparatus (referred to as a prior art) will be described with reference to FIG. To do. 4A and 4B, it is assumed that pictures are displayed in the order of P0, B1, B2, P3, and B4. Here, P0 and P3 indicate unidirectional predictive encoded images (P pictures), B1, B2, and B4 indicate bidirectional predictive encoded images (B pictures), and the numbers after P and B are displayed. Indicates the order. It is assumed that a scene change has occurred between the picture B2 and the picture P3.

  In this case, when an error occurs in the picture B2, which is a non-reference picture immediately before the sea change, in the prior art, as shown in FIG. 4B, there are a plurality of reference pictures before and after the error picture B2. It is determined that the picture has a MV value that complements the pictures P0 and P3. However, the scene changes between the picture B2 and the picture P3, and the MV value of the picture P3 is not an appropriate MV value predicted purely from the picture P0.

  In contrast, in the present embodiment, when an error occurs in picture B2, which is a non-reference picture, each of pictures P0 and P3, which are reference pictures immediately before and after error picture B2, is occupied in the picture. The intra MB ratio is compared with the threshold value F to determine the presence or absence of correlation, and the average value and variance value of the MV values of each MB of the picture P3 and the picture P0 are calculated. In this case, the average value of the MV values When the variance value is calculated, the average value and the variance value of the picture P0 of the same scene as the error picture B2 are smaller than the average value and the variance value of the picture P3 after the scene change occurs, and thus the picture having the most correlation As shown in FIG. 4A, only the picture P0 is selected as the picture that complements the MV value. Can.

  The present invention includes a program for causing a computer to realize the functions of the above apparatus. This program may be read from a recording medium and loaded into a computer, or may be transmitted via a communication network and loaded into a computer.

  In the above embodiment, the I picture and the P picture are described as reference pictures and the B picture is described as a non-reference picture. However, the present invention is not limited to this, and an encoding standard (for example, H.264) is used. Depending on the H.264 standard), even a B picture can be a reference picture, and a P picture can be set as a non-reference picture. Further, when the error picture is a non-reference picture, the reference pictures having MV values before and after the error picture have been described as the complement target pictures. However, depending on the encoding standard (for example, the H.264 standard), two reference pictures in the forward direction are used. The picture may be the complement target picture, the two backward reference pictures may be the complement target pictures, and is not limited to the immediately preceding or immediately following reference picture.

  This is effective for a moving picture decoding apparatus used in a communication device, a mobile device or the like that requires error processing.

It is a block diagram of one embodiment of the device of the present invention. It is a flowchart for general | schematic operation | movement description of this invention apparatus. It is a flowchart for operation | movement description of the correlation judgment part in FIG. It is a figure explaining the difference of one embodiment of this invention and the complementary picture of a prior art. It is a block diagram of an example of a conventional device. It is a MV complementation figure of MB unit of patent documents 1. 10 is a flowchart related to MV complementation in Patent Document 1;

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Input bit stream 102 Memory 103 Decoding part 104 Inverse quantization part 105 Inverse conversion part 106 Decoded image output 107 Adder 108 Error detection part 109 MV value calculation part 110 MC part 111 Frame memory 112 Decoding order control part 113 Correlation judgment part 114 MV value storage

Claims (2)

Each picture of the video signal is divided into block units that are image areas of a predetermined number of pixels, and the encoded video data generated by using orthogonal transform, quantization, and variable length coding for each block unit. A non-reference picture to be decoded with reference to the motion vector information of another picture, or the motion vector information of the picture itself A moving image decoding apparatus that generates decoded image data by performing decoding, inverse quantization, and inverse orthogonal transform on moving image encoded data including a reference picture that is referred to when decoding a picture. ,
Decoding means for decoding the moving image encoded data and outputting decoded coefficient data and decoded MV data;
Error detecting means for detecting whether an error has occurred during decoding of the moving image encoded data decoded by the decoding means;
When an error is detected by the error detection unit, a decoding order control unit that controls a decoding order of the moving image encoded data according to a picture type of the error in which the error is detected;
MV value storing means for holding the decoded MV data output from the decoding means in accordance with a control signal output from the decoding order control means by controlling the decoding order;
MV calculation means for calculating an MV value using at least MV data from the MV value storage means and the decoded MV data from the decoding means;
When the picture in which the error is detected by the error detection means is the non-reference picture, one or more reference pictures on at least one of the front and rear sides in the decoding order of the non-reference picture are set as the complement target pictures. A correlation determination means for determining the presence or absence of correlation by comparing the presence ratio of the intra block in the complement target picture with a predetermined threshold;
When there is not one complement target picture determined to be correlated by the correlation determination unit, at least one of an average value and a variance value of MV values from the MV calculation unit for each of the plurality of complement target pictures Is calculated as a picture having a complementary MV value, and when the number of the complementary pictures determined to be correlated by the correlation determining means is one, Complementary picture determining means for determining a complementary target picture as a picture having a complementary MV value;
When the image data obtained by performing the inverse quantization and the inverse transform on the decoded coefficient data output from the decoding means is the non-reference picture, the decoded reference picture data and the Decoded image data for generating the decoded image data by adding the motion compensated image data generated from the MV value from the MV value calculating means, and storing the decoded image data of the reference picture as the decoded reference picture data Generating means;
When the decoded image data generating means generates decoded image data of a picture in which an error has been detected by the error detecting means, decoded MV data of a picture having the complementary MV value determined by the complementary picture determining means MV data complementing means for supplementing the MV value as the MV value from the MV value calculating means. Each picture of the video signal is divided into block units that are image areas of a predetermined number of pixels, and the encoded video data generated by using orthogonal transform, quantization, and variable length coding for each block unit. A non-reference picture to be decoded with reference to the motion vector information of another picture, or the motion vector information of the picture itself Using a computer, decoding that generates decoded image data by performing decoding, inverse quantization, and inverse orthogonal transform on moving image encoded data including a reference picture that is referred to when decoding a picture is performed. A moving picture decoding program to be realized,
The computer,
Decoding means for decoding the moving image encoded data and outputting decoded coefficient data and decoded MV data;
Error detecting means for detecting whether an error has occurred during decoding of the moving image encoded data decoded by the decoding means;
When an error is detected by the error detection unit, a decoding order control unit that controls a decoding order of the moving image encoded data according to a picture type of the error in which the error is detected;
MV value storing means for holding the decoded MV data output from the decoding means in accordance with a control signal output from the decoding order control means by controlling the decoding order;
MV calculation means for calculating an MV value using at least MV data from the MV value storage means and the decoded MV data from the decoding means;
When the picture in which the error is detected by the error detection means is the non-reference picture, one or more reference pictures on at least one of the front and rear sides in the decoding order of the non-reference picture are set as the complement target pictures. A correlation determination means for determining the presence or absence of correlation by comparing the presence ratio of the intra block in the complement target picture with a predetermined threshold;
When there is not one complement target picture determined to be correlated by the correlation determination unit, at least one of an average value and a variance value of MV values from the MV calculation unit for each of the plurality of complement target pictures Is calculated as a picture having a complementary MV value, and when the number of the complementary pictures determined to be correlated by the correlation determining means is one, Complementary picture determining means for determining a complementary picture as a picture having a complementary MV value;
When the image data obtained by performing the inverse quantization and the inverse transform on the decoded coefficient data output from the decoding means is the non-reference picture, the decoded reference picture data and the Decoded image data for generating the decoded image data by adding the motion compensated image data generated from the MV value from the MV value calculating means, and storing the decoded image data of the reference picture as the decoded reference picture data Generating means;
When the decoded image data generating means generates decoded image data of a picture in which an error has been detected by the error detecting means, decoded MV data of a picture having the complementary MV value determined by the complementary picture determining means Is decoded as an MV data complementing unit that complements the MV value from the MV value calculating unit.

Images