JPH09271025A - Error processing unit - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an excellent display image by conducting skip read processing depending on a layer on the occurrence of an error so as to conduct efficient decoding processing even in the case of occurrence of an error in the decoding processing of the MPEG. SOLUTION: A header detection section 10 detects a start course and an end course in a header of a bit stream, and on the occurrence of an error in a variable length decoding section 12, prescribed read skip processing is conducted. Furthermore, the variable length decoding section 12 decodes a variable length code and detects a data error in the bit stream. In the decoding processing of the bit stream, on the occurrence of an error, skip reading is conducted up to a decoding enable position depending on a layer of error occurrence. Thus, it is not required to stop decoding processing and efficient decode processing is conducted. Since substitute data are displayed in place of skipped data, a visually excellent display image is obtained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an error processing device in image decoding processing, and more particularly to MPE.
The present invention relates to an error processing device in a G decoding process.

[0002]

2. Description of the Related Art Conventionally, MPEG has been known as a standard image coding technique for mainly compressing moving image data. Here, the image data in MPEG is sequence G
It has a six-layer hierarchical structure of OP (group of pictures), picture, slice, macroblock (MB), and block.

In the MPEG bit stream, a header is attached to the beginning of each of the above-mentioned sequences, GOPs, pictures and slices, which are constituent units. That is, as shown in FIG. 14, a sequence has a sequence header at the front position of each sequence data, and a GOP has a GOP header at the front position of each GOP data. Also, for pictures, a picture header is provided at the front position of each picture data, and for slices, a slice header is provided at the front position of each slice data. A sequence end code indicating the end of each sequence is attached to the end of each sequence.

Further, a sequence header code (SHC) is added to the beginning of the sequence header,
A GOP start code (GSC) is added to the head of the GOP header. A picture start code (PSC) is attached at the head of the picture header, and a slice start code (SSC) is attached at the head of the slice header.

[0005]

However, in the decoding processing, what kind of processing is to be performed when a data error occurs in the bit stream is not specified in the MPEG standard. In addition, when the above error occurs, the decoding process may be stopped, but in that case, the decoding process thereafter cannot be performed due to some errors, and the efficient decoding process may be performed. There is a problem that it is not possible, and it is desired to obtain a visually good display image even when an error occurs. Therefore, an object of the present invention is to provide an error processing device capable of obtaining a visually good display image by performing efficient decoding processing even when an error occurs in the decoding processing in MPEG. To do.

[0006]

The present invention was created to solve the above problems. First, it performs error processing when decoding image data having a hierarchical structure. The error processing device is characterized by performing a skip processing according to a type of a layer in which an error has occurred.
According to the error processing device of the first configuration, the skip processing is performed according to the layer in which the error has occurred, so that efficient decoding processing can be performed without stopping the decoding processing. Further, since the skip processing is performed according to the type of hierarchy, it is only necessary to provide processing according to the number of types of hierarchy, and since it is only necessary to provide error processing for the number of types of hierarchy, Processing becomes easy.

Secondly, there is provided an error processing device for performing an error process when decoding image data having a hierarchical structure, which is at least a header detection part, a variable length decoding part, and an inverse quantization. Section and an inverse two-dimensional DCT transform section,
The variable length decoding unit detects an error that has occurred in each layer, and the header detection unit, in order to perform the skip processing according to the layer in which the error occurs, the start code or header code of a predetermined layer. Characterized by searching. In the error processing device having the second configuration, first, the header detection unit detects the start code or the header code of each layer, and the variable length decoding unit performs the header process of each layer. Then, when an error is detected, the skip processing is performed up to a predetermined layer. Therefore, efficient decoding processing can be performed without stopping the decoding processing.

Thirdly, if an error occurs in the processing of the sequence header of the sequence in the image data, the reading is skipped to the next sequence. That is, when an error occurs in the header processing at the sequence level, the appropriate decoding processing cannot be performed until the next sequence, and thus the next sequence is skipped. As a result, efficient decoding processing can be performed. Fourthly, when an error is detected in the processing of the GOP header of the GOP of the image data, or when an error is detected in the processing of the picture header of the I picture or P picture in the picture, the next layer is added. Yes, I picture hierarchy or G
It is characterized in that reading is skipped up to a hierarchy of GOP or higher including OP. That is, when an error occurs in the header processing of the GOP header of the GOP or the header processing of the picture header of the picture excluding the B picture, the appropriate decoding processing is performed up to the next intra picture or up to the layers of GOP and above. Since this is not possible, the next I picture or GOP or higher is skipped. As a result, efficient decoding processing can be performed.

Fifthly, when an error is detected in the processing of the picture header of the B picture in the picture of the image data, the next layer, which is the layer of one of the pictures or a GOP including a GOP or more. It is characterized by skipping to the hierarchy. In other words, when an error occurs in the header processing of a B picture, it is not possible to perform appropriate decoding processing up to the next picture or layers up to GOP or higher, so the next picture or GO
Skips to P or higher layers. As a result, efficient decoding processing can be performed. Further, sixthly, when an error is detected in the processing of the slice header in the slice of the image data or the processing of each macroblock, the next layer, up to the layer of the slice or a layer including a picture or more including the picture is detected. Characterized by skipping. That is, when an error occurs in the slice processing, appropriate decoding processing cannot be performed up to the layer including the next slice or higher, and therefore, skipping is performed to the layer higher than the next slice.

A seventh feature is that a substitute display image is displayed for the image data in the skipped range. By displaying an appropriate display image in this way, a visually good display image can be obtained. Eighthly, a display image immediately before the skipped range is displayed as a substitute display image. Thereby, the continuity of images can be maintained. Ninth, a black image is displayed as a substitute display image. As a result, the viewer does not feel uncomfortable as compared with the case of displaying an image that has not been properly decoded. The tenth feature is that a display image displaying a message is displayed as a substitute display image. This allows the viewer to be notified of a message such as "please wait for a while".

The eleventh feature is that the image data in the skipped range is interpolated with a substitute image for each macroblock unit. As a result, interpolation is performed for the skipped slices, and interpolation is performed with an appropriate image, so that a visually good display image can be obtained. The twelfth feature is that interpolation is performed using an image cut out and extracted from a reference picture as a substitute image. This makes it possible to maintain image continuity. In addition, thirteenth, as a substitute image,
It is characterized in that interpolation is performed by a black display block. As a result, the viewer does not feel uncomfortable as compared with the case where an image that has not been properly decoded is displayed.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described with reference to the drawings. A decoding device T as an error processing device according to the present invention, as shown in FIG.
A header detection unit 10, a variable length decoding unit 12, an inverse quantization unit 14, an inverse two-dimensional DCT conversion unit 16, a reference MB cutout unit 18, a frame memory A, a frame memory B,
Frame memory C, switch SW1, switch SW
2, a switch SW3, and a processing control unit 20.

Here, the header detecting section 10 detects a start code and an end code in a header in a bit stream, and when an error is detected in the variable length decoding section 12, a predetermined skip is skipped. Perform processing. Further, the variable length decoding unit 12 decodes the variable length code and detects a data error in the bitstream. For example, the error in the data following the start code in the header is detected, and the error in the macroblock processing in the slice data is detected.

Further, the inverse quantization section 14 performs inverse quantization processing on the bit stream. That is, a process of multiplying the quantization value of the bit stream by a quantization step is performed to calculate a DCT coefficient. The inverse two-dimensional DCT transform unit 16 also performs DCT inverse transform processing on the DCT coefficient calculated by the inverse quantization unit 14. Further, the reference MB cutout unit 18 cuts out and extracts the image block specified by the motion vector from the reference frame. The processing control unit 20 controls the operation of each unit. In addition, the frame memory A,
The frame memories B and C store image data of a certain frame.

The switch SW1 switches between four input terminals and one output terminal, and the first terminal of the four input terminals is connected to the output side of the inverse two-dimensional DCT conversion section 16, The second terminal is the inverse two-dimensional D
It is connected to the terminal connecting the output side of the CT conversion section 16 and the output side of the reference MB cutout section 18, the third terminal is connected to the output side of the reference MB cutout section 18, and the fourth terminal is grounded. It is grounded. The output terminal of the switch SW1 is connected to the input terminal of the switch SW2. The first terminal to the fourth terminal are shown by the numbers attached to the switch SW1 in FIG.

The switch SW2 switches between one input terminal and three output terminals. The first terminal of the three output terminals is connected to the frame memory C and the second terminal is connected to the frame memory A. , And the third terminal is connected to the frame memory B. It should be noted that the first to third terminals are the switches S in FIG.
It is shown by the number attached to W2.

The switch SW3 switches between four input terminals and one output terminal. The first terminal of the four input terminals is connected to the frame memory C and the second terminal is connected to the frame memory A. Connected to the third
The terminal is connected to the frame memory B, and the fourth terminal is grounded. Further, the output terminal of the switch SW3 is an image data output terminal of the decoding device T. The first terminal to the fourth terminal are indicated by the numbers attached to the switch SW3 in FIG.

The operation of the decoding device T having the above configuration will be described. First, a description of the normal operation, that is, a certain GOP
In the processing of 1, the operation when no error occurs will be described. The sequence of pictures in the original image is "B0-B1-I
2-B3-B4-P5-B6-B7-P8 ". Here, the first alphabetic character represents the picture type. For example, B0 indicates a B picture and I2 indicates I.
A picture is shown, and P5 is a P picture.
It is assumed that each picture is a sequence of pictures in a certain GOP having a certain sequence. In the case where the pictures in the original image are arranged as described above, in relation to the prediction direction,
The sequence of pictures on the media is "I2-B0-B1.
-P5-B3-B4-P8-B6-B7 ", and the decoding process is performed in this order.

To explain the specific processing, first,
The picture I2 is decoded. That is, in the header detection unit 10, the header detection unit 10 detects the start code (header code in the case of a sequence) of each header in order of sequence, GOP, and picture, and in the variable length decoding unit 12, each header is detected. Header processing is performed. Also for the slice, the header detection unit 10 detects the slice start code of each slice, and the variable length decoding unit 12 performs slice processing. That is, slice header processing and macroblock processing are performed. This slice processing is performed for each slice of a picture. Further, the inverse quantization unit 14 performs inverse quantization processing,
The inverse two-dimensional DCT transform unit 16 performs the inverse DCT transform.

In this case, the switch SW1 is switched to the first terminal side, the switch SW2 is switched to the second terminal side, and the image data of the picture I2 is stored in the frame memory A. The switch SW3 is switched to the side of the fourth terminal, and nothing is displayed during this picture I2 processing.

Next, the picture B0 is processed. That is, in this case, after the start code of the picture header is detected and the header process is performed, the start code of the slice header and the slice process are sequentially performed for each slice. The procedure of this process is
The same applies to the processing of each picture below.

In the case of processing this picture B0, when the macro block in the slice is an intra macro block, the switch SW1 is on the first terminal side, and the switch SW2.
Is on the first terminal side, and the switch SW3 is on the first terminal side, and the decoded image data is output from the decoding device T. When the switch SW2 is on the first terminal side, the image data is once stored in the frame memory C and output to the switch SW3 side. On the other hand, if the macroblock is an inter macroblock and prediction is performed from the picture I2, the switch SW1 is on the second terminal side, the switch SW2 is on the first terminal side,
The switch SW3 is on the first terminal side, and the frame memory A
The decoded image is output from the decoding device T by referring to the image data of the picture I2 stored in the. That is, when processing this picture B0, the picture B0 is output.

Next, the picture B1 is processed. In this case, too, the same processing as in the case of the picture B0 is carried out.
When the macroblock in the slice is an intra macroblock, the switch SW1 is on the first terminal side, the switch SW2 is on the first terminal side, and the switch SW3 is on the first terminal side, and the decoded image data is output from the decoding device T. On the other hand, if the macroblock is an inter macroblock and the prediction is performed from the picture I2, the switch SW1 is on the second terminal side, the switch SW2 is on the first terminal side, and the switch S is
W3 becomes the first terminal side, and the decoded image data referring to the image data of the picture I2 stored in the frame memory A is output from the decoding device T. That is, this picture B
At the time of processing 1, the picture B1 is output.

Next, the picture P5 is processed. Here, when the macroblock in the slice is an intra macroblock, the switch SW1 is on the first terminal side, the switch SW2 is on the third terminal side, and the image data of the picture P5 is stored in the frame memory B. On the other hand, if the macroblock is an inter macroblock and the prediction is performed from the picture I2, the switch SW1 is on the second terminal side and the switch SW2 is on the third terminal side, and the image data of the picture I2 stored in the frame memory A is The referred image data is stored in the frame memory B. further,
In the case of both intra macro blocks and inter macro blocks, the switch SW3 is on the second terminal side,
The decoded image data of the picture I2 is output.

Next, the picture B3 is processed. In this case as well, the same processing as in the case of the B picture is carried out. If the macroblock in the slice is an intra macroblock, the switch SW1 is on the first terminal side. , The switch SW2 is on the first terminal side, the switch SW3 is on the first terminal side, and the decoded image data is output from the decoding device T. On the other hand, if the macroblock is an inter macroblock and the prediction is performed from the pictures I2 and P5, the switch S
W1 is on the second terminal side, switch SW2 is on the first terminal side, and switch SW3 is on the first terminal side, and the image data of the picture I2 stored in the frame memory A and the picture P5 stored in the frame memory B is referred to. The decoded image data is output from the decoding device T. That is, this picture B
At the time of processing 3, the picture B3 is output.

Next, the process of the picture B4 is performed, and in this case, the same process as that of the picture B3 is performed.
That is, when the macroblock in the slice is an intra macroblock, it is output from the decoding device T as it is. On the other hand, when the macroblock is an inter macroblock and is predicted from the pictures I2 and P5, it is stored in the frame memory A. Decoded image data that refers to the image data of the picture I2 that was created or the picture P5 that is stored in the frame memory B is output from the decoding device T. That is, at the time of processing this picture B4, the picture B4 is output.

Next, the picture P8 is processed. Here, when the macroblock in the slice is an intra macroblock, the switch SW1 is on the first terminal side, the switch SW2 is on the second terminal side, and the image data of the picture P8 is stored in the frame memory A. On the other hand, assuming that the macro block is an inter macro block and the prediction is performed from the picture P5, the switch SW1 is on the second terminal side and the switch SW2 is on the second terminal side, and the image data of the picture P5 stored in the frame memory B is displayed. The referred image data is stored in the frame memory A. further,
In the case of both the intra macro block and the inter macro block, the switch SW3 is on the third terminal side,
The image data of the picture P5 is output.

Next, the picture B6 is processed. In this case as well, the same processing as in the case of the B picture is carried out. If the macroblock in the slice is an intra macroblock, the switch SW1 is on the first terminal side. , The switch SW2 is on the first terminal side, the switch SW3 is on the first terminal side, and the decoded image data is output from the decoding device T. On the other hand, if the macroblock is an inter macroblock and the prediction is performed from the pictures P5 and P8, the switch S
W1 is on the second terminal side, switch SW2 is on the first terminal side, and switch SW3 is on the first terminal side, and the image data of the picture P8 stored in the frame memory A and the picture P5 stored in the frame memory B is referred to. The decoded image data is output from the decoding device T. That is, this picture B
At the time of processing 6, the picture B6 is output.

Also in the next processing of the picture B7, the same processing as in the case of the picture B3 is performed. That is, when the macroblock in the slice is an intra macroblock, it is output from the decoding device T as it is. On the other hand, when the macroblock is an inter macroblock and is predicted from the pictures P5 and P8, it is stored in the frame memory A. The decoding device T outputs the image data that refers to the image data of the picture P8 that is stored and the image data of the picture P5 that is stored in the frame memory B. That is,
At the time of processing this picture B7, the picture B7 is output.

Next, an error processing method in the decoding device T of this embodiment will be described. That is, when an error is detected in the header processing in the sequence, GOP, picture, and slice, or in the macroblock processing of each slice, the skip processing is performed to perform the error processing. This will be specifically described below.

First, as shown in FIG. 2, a header search is performed (S101). Here, a sequence level search is performed. Here, the sequence level search is
Skipping is performed until a sequence header code or sequence end code is detected, and skipping is performed even if a start code in a layer below a picture including a picture is detected. This header search is performed by the header detection unit 1
0. Then, if the detected code is a sequence end code, the processing is terminated (S10).
2, S103), in the case of a sequence header code,
Sequence header processing is performed (S104, S105).
If an error is detected in this header processing, error processing 1 is performed (S106, S107). Here, if an error is detected before the start code of the GOP header is detected, this error processing 1 is performed. As an error case, there is a case where nonstandard data is detected, including the following error cases. The detection of each error including the error processing 1 is performed by the variable length decoding unit 12.

In the error processing 1, the reading is skipped, but at this time, the search at the sequence level or higher is performed. This search is performed by the header detection unit 10, and, for example, as shown in FIGS. 6 and 7, skips reading until the next sequence header code is detected. In FIG. 7, the Nth sequence indicated by hatching is the skipped sequence. The header detection unit 10 does not transmit the skipped image data to the variable length decoding unit 12. Then, as the display of the skipped range, the following cases can be considered.

First, as the first method, the immediately preceding image is displayed. That is, when there is a previous display image, this previous image data is stored in any of the frame memory A, the frame memory B, and the frame memory C, and therefore this image data is output. According to this first method, since the immediately preceding image is displayed, the continuity of the images can be maintained. As a second method, black screen display is performed. That is, black screen display is performed with the input terminal of the switch SW3 set to the fourth terminal side. According to the image display of the second method, the black screen display is performed, so that the viewer does not feel uncomfortable. In other words, when displaying an image that has not been properly decoded without skipping, there is a high possibility that an image that is difficult to identify will be displayed. It does not give a pleasant feeling. Also,
As a third method, displaying a message can be considered. That is, the fourth terminal of the switch SW3 is connected to the output side of the message information, and a message such as "please wait for a while" is displayed. According to this method, it is possible to notify the viewer of the message, and it is possible to notify that it is a trouble on the side of the information source such as broadcast and not a hardware trouble.

As described above, when an error occurs in the header processing at the sequence level, the appropriate decoding processing cannot be performed until the next sequence. Therefore, the next sequence layer is skipped and skipped. For the data, an alternative image is displayed. If the searched code is not the sequence header code in step S104, the process returns to step S101.

Next, if no error is detected in step S106, a header search is performed as shown in FIG. 3 (S108). Here, a search at a picture level or higher is performed. Here, the search at the picture level or higher means skipping until one of the sequence header code, sequence end code, group start code, and picture start code is detected, and the slice start code is detected. Ignore it and skip over it. Then, it is determined whether or not the searched code is the group start code in the GOP, and if it is the group start code, the GOP header processing is performed (S109, S110). When an error is detected in this header processing, error processing 2 is performed (S111, S112). In other words, if an error is detected before the start code or header code of the header of the picture level or higher, this error processing 2 is performed. For example, if an error is detected before the picture start code of the first I picture is detected, this error processing will be performed.

In this error processing 2 as well, skipping is performed, but at that time, a search at a picture level or higher is performed,
In the case of pictures, it is limited to I pictures. Therefore, if any of the I picture start code, sequence header code, or sequence end code is searched, the skip processing is stopped at that point, but even if the P picture or B picture header code is detected, those Ignore the picture and skip it. This search is performed by the header detection unit 10. The header detection unit 10 does not transmit the skipped image data to the variable length decoding unit 12. Then, as the display of the skipped range, similar to the case of the error processing 1, the display of the immediately preceding image, the display of the black screen, the display of the message, and the like are performed. If no error is detected in step S111, step S111
Return to 118.

Further, in the determination in step S109,
If the searched code is not a group start code, it is determined whether or not it is a picture start code (S
113), in the case of a picture start code, the picture type is detected (S114). The variable length decoding unit 12 detects the picture type. On the other hand, step S
If the code is not the picture start code in 113, the process returns to step S102.

In the detection of step S114, it is determined whether or not the picture type is an I picture (S11).
5) If it is not an I picture, the above error processing 2 is performed. That is, since the first picture to be decoded in a certain GOP should be an I picture, if the picture type of the I picture cannot be detected, it is judged as an error, and if it is a picture at the picture level or higher, a picture is detected. Skips over to the I picture and skips over to the next I picture or GOP. In addition,
In determining the picture type including step S114, the variable length decoding unit 12 determines from the data following the picture start code in the picture header.

If it is determined in step S115 that the picture type is I picture, I picture
Picture header processing is performed (S116). It is determined whether or not an error is detected in this header processing (S
117), if an error is detected, error processing 2 is performed (S112). That is, if an error is detected in the header processing before the start code of the header of a layer higher than the slice level is detected, this error processing 2 is performed. For example, if an error is detected before the slice start code of the first slice is detected, this error processing will be performed.

Next, if no error is detected in step S117, header search is performed as shown in FIG. 5 (S128). Here, the search is performed at the slice level or higher. Here, the search at the slice level or higher means skipping reading until any of the sequence header code, the sequence end code, the group start code, the picture start code, and the slice start code is detected. Then, it is determined whether or not the searched code is a slice start code (S129), and if it is a slice start code, slice processing is performed (S130). That is, the slice header process is performed to detect the quantization width q, and further the macroblock process is performed to decode the actual image data.

If an error is detected in this slice processing, error processing 4 is performed (S131, S13).
2). That is, when an error is detected in the slice processing before the start code or header code of the header of the slice level or higher is detected, this error processing 4 is performed. For example, if an error is detected before the slice start code of the next slice is detected, this error processing will be performed.

In this error processing 4 as well, reading is skipped, but at this time, a search at a slice level or higher is performed.
This search is performed by the header detection unit 10, and, for example, as shown in FIG.
2. As shown in FIG. 13, skipping is performed until the next slice start code. In FIG. 13, the third slice indicated by hatching is the skipped slice. The header detection unit 10 does not transmit the skipped image data to the variable length decoding unit 12. Then, as the display of the skipped range, interpolation is performed by an alternative process in units of macroblocks. For example, interpolation is performed by extracting and extracting a reference picture such as the immediately preceding reference frame from the same position, or interpolation using a concealment motion vector. Further, the black display block may be used for interpolation.

In the error processing 4, when an error occurs in the slice processing, the appropriate decoding processing cannot be performed up to the layer including the next slice, so that the layers up to and including the next slice are read. For the skipped and skipped data, a substitute image is displayed.

Next, in step S129, if it is determined that the slice start code is not detected, the process proceeds to step S118. As a case where the step S129 transits to the step S118, it is conceivable that the slice processing of a certain picture is all finished and the processing proceeds to the next picture.

In step S118, it is determined whether or not the searched code is the picture start code, and if it is the picture start code, the picture type is detected (S119). Then, when the detected picture type is P picture, P picture header processing is performed (S121). It is determined whether or not an error is detected in this header processing (S122), and when an error is detected, error processing 2 is performed (S112).
That is, if an error is detected in the header processing before the start code of the header of a layer higher than the slice level is detected, this error processing 2 is performed. For example,
If an error is detected before the slice start code of the first slice is detected, this error processing will be performed. In this error processing 2, skipping is performed similarly to the error processing 2 described above, and for example, as shown in FIGS. 8 and 9, skipping is performed up to the start code of the next I picture. In FIG. 9, the P-picture to the B-picture shown by hatching are skipped pictures.

As described above, in the error processing 2, the GOP
When an error occurs in the header processing of the GOP header of the above, or the header processing of the picture header of the picture excluding the B picture, until the next intra picture, or GO
Since an appropriate decoding process cannot be performed up to P or higher layers, the next I picture or GOP or higher layers are skipped, and a substitute image is displayed for the skipped data. If no error is detected in step S122, the process proceeds to step S128.

On the other hand, if the picture type detected in step S119 is a B picture, B picture header processing is performed (S124). It is determined whether or not an error is detected in this header processing (S125), and if an error is detected, error processing 3 is performed (S11).
2). That is, if an error is detected in the header processing before the start code of the header of a layer higher than the slice level is detected, this error processing 3 is performed. For example, if an error is detected before the slice start code of the first slice is detected, this error processing will be performed.

In this error processing 3 as well, skipping is performed, but at that time, a search at a picture level or higher is performed.
Therefore, if any of the picture start code, sequence header code, and sequence end code is searched, the skip processing is stopped. In the case of the picture start code, there is no limitation on the picture type as in the error processing 2 above. This search is performed by the header detection unit 10, and, for example, as shown in FIGS. 10 and 11, skipping is performed up to the start code of the next B picture. FIG.
In, the B picture indicated by hatching is a skipped picture. The header detection unit 10 does not transmit the skipped image data to the variable length decoding unit 12. Then, as the display of the skipped range, the display of the immediately preceding image, the display of the black screen, the display of the message and the like are performed as in the case of the error processing 1 and the error processing 2.
If no error is detected in step S125, the process returns to step S127. Step S119
In the detection of the picture type in, if it is neither a P picture nor a B picture, the process proceeds to step S109.

In this error processing 3, when an error occurs in the header processing of the B picture, the appropriate decoding processing cannot be performed up to the next picture or up to a layer higher than GOP, so that the next picture or It skips to a level higher than GOP and displays a substitute image for the skipped data.

From the above, regarding the error in the picture, the above error processing 2 is performed in the case of the I picture and the P picture, and the above error processing 3 is performed in the case of the B picture.

As described above, according to the decoding apparatus T of the present embodiment, when an error occurs in the bitstream decoding process, the read operation is skipped to a decodable position according to the layer in which the error occurred. Since this is performed, it is possible to perform efficient decoding processing without having to stop the decoding processing. Further, for the data skipped, a substitute image is displayed, so that a visually excellent display image can be obtained. Further, according to the decoding apparatus T of the present embodiment, the skip processing according to the type of hierarchy is performed, so that the processing according to the number of types of hierarchy may be provided. Also, skip processing is performed according to the type of hierarchy, and the same processing is performed for any data if an error occurs in a certain hierarchy. Therefore, if the number of error handling methods is set according to the number of types of hierarchy. Since it is good, the processing becomes easy.

[0052]

According to the error processing device of the present invention, the skip processing according to the type of the layer in which the error has occurred is performed, and the image data in the skipped range is displayed as a substitute display image. By displaying or by interpolating with a substitute image for each macroblock unit, efficient decoding processing can be performed without stopping the decoding processing.

[Brief description of drawings]

FIG. 1 is a block diagram illustrating a configuration of a decoding device based on an embodiment of the present invention.

FIG. 2 is a flowchart showing an operation of the decoding device according to the embodiment of the present invention.

FIG. 3 is a flowchart showing an operation of the decoding device according to the embodiment of the present invention.

FIG. 4 is a flowchart showing an operation of the decoding device according to the embodiment of the present invention.

FIG. 5 is a flowchart showing an operation of the decoding device according to the embodiment of the present invention.

FIG. 6 is an explanatory diagram illustrating error processing 1.

FIG. 7 is an explanatory diagram illustrating error processing 1;

FIG. 8 is an explanatory diagram illustrating error processing 2;

FIG. 9 is an explanatory diagram illustrating error processing 2;

FIG. 10 is an explanatory diagram illustrating error processing 3;

FIG. 11 is an explanatory diagram illustrating error processing 3;

FIG. 12 is an explanatory diagram illustrating error processing 4;

FIG. 13 is an explanatory diagram illustrating an error process 4.

FIG. 14 is an explanatory diagram showing the structure of image data in MPEG.

[Explanation of symbols]

 T Decoding device 10 Header detection unit 12 Variable length decoding unit 14 Inverse quantization unit 16 Inverse two-dimensional DCT conversion unit 18 Reference MB cutout unit 20 Processing control unit

Continued Front Page (72) Inventor Anpachi Hamamoto 2-5-5 Keihan Hondori, Moriguchi City, Osaka Sanyo Electric Co., Ltd.

Claims (13)

[Claims] 1. An error processing device for performing an error process when decoding image data having a hierarchical structure, characterized by performing a skip process according to a type of a layer in which an error has occurred. Error handling device. 2. An error processing device for performing an error process when decoding image data having a hierarchical structure, comprising at least a header detection unit, a variable length decoding unit, an inverse quantization unit, and an inverse unit 2. A variable dimension decoding unit, the variable length decoding unit detects an error occurring in each layer, and the header detecting unit performs a predetermined skipping process according to the layer in which the error occurs. An error processing device characterized by searching for a start code or a header code of the hierarchy. 3. When an error is detected in the processing of the sequence header of the sequence in the image data,
The error processing device according to claim 1 or 2, wherein reading is skipped up to the next sequence. 4. In the next layer, when an error is detected in the processing of the GOP header of the GOP of the image data, or when an error is detected in the processing of the picture header of the I picture or P picture in the picture. 4. The skipping is performed up to the IOP layer, the I picture layer, or the GOP layer including the GOP or higher layers.
The error processing device according to any one of 1. 5. When an error is detected in the processing of the picture header of the B picture in the picture of the image data, the next layer, which is the layer of any picture or the GOP including GOP or higher, is skipped. The error processing device according to claim 1, wherein the error processing device is configured to perform the following. 6. When an error is detected in the processing of the slice header in the slice of the image data or the processing of each macroblock, the next layer, which is the layer of the slice or a layer including a picture or more including a picture, is skipped. The error processing device according to any one of claims 1 to 5, wherein the error processing device is configured to perform the following. 7. The error processing apparatus according to claim 1, wherein a substitute display image is displayed for the image data in the skipped range. 8. The error processing device according to claim 7, wherein a display image immediately before a range of skipped reading is displayed as a substitute display image. 9. The error processing device according to claim 7, wherein a black image is displayed as a substitute display image. 10. The display image displaying a message is displayed as a substitute display image.
The error processing device described in. 11. The error processing apparatus according to claim 6, wherein the image data in the skipped range is interpolated with a substitute image for each macroblock unit. 12. The error processing apparatus according to claim 11, wherein interpolation is performed using an image extracted and extracted from a reference picture as a substitute image. 13. The error processing device according to claim 11, wherein the substitute image is interpolated by a black display block.