JP2003116129A - Image compression data decoder - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image compression data decoder capable of easily writing image data subjected to decoding and resolution conversion to a display memory and decreasing disturbance in a decoded image caused on the occurrence of a decoding error of a variable length code. SOLUTION: In order to reduce the memory capacity, image data are stored in a display memory 16 after reducing the image data by using resolution conversion resulting in reducing the memory capacity. The decoder is provided with a macro block number output section 14 and a display memory write control section 15 to write the image data on the way of a line buffer in the unit of macro blocks and is configured not to overwrite ineffective data to data of the display memory equivalent to a block earlier than an erroneous packet in the display memory thereby making disturbance of a decoded image unremarkable with the memory capacity as small as possible on the occurrence of a decoding error of the variable length code.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a compressed image data decoding apparatus for decoding compressed image data, converting the resolution of the decoded image data, and displaying the converted image data.

[0002]

2. Description of the Related Art In the JPEG, MPEG1, MPEG2, and MPEG4 standards, which are known as image compression encoding methods, the discrete cosine transform (DCT) of still image data or moving image data is quantized, and the obtained quantization is obtained. The DCT coefficient is further subjected to variable length coding using Huffman coding or the like to obtain compressed data.

When the decoding device for decoding the image data compressed in this way is constituted by an LSI, the decoding LSI is a variable-length code decoding processing section for decoding a variable-length code on the compressed image data, and is decoded. An inverse quantization unit that inversely quantizes data, and an inverse discrete cosine transformation unit that obtains image data by performing an inverse discrete cosine transform on the inverse quantized DCT coefficient,
Further, in the case of MPEG1, MPEG2, and MPEG4, since moving image compression is performed, image data further including a motion compensating unit for performing motion compensation on image data obtained by inverse discrete cosine transform. It has a decryption unit.

Here, the variable-length code decoding unit of the decoding LSI has a variable-length code table such as Huffman code, and the variable-length code is cut out from the bit string of the variable-length code by referring to this table and the corresponding run code is extracted.・ Level output,
That is, the decrypted data is obtained. In the run-level output, the level is the quantized D
It is a DCT coefficient which is not “0” in the code bit string obtained by zigzag scanning the CT coefficient, and the run (ru)
n) represents the number of consecutive "0" s preceding the above level.

When the variable length code decoding unit refers to the variable length code table and the variable length code corresponding to the code bit string is not entered, a variable length code decoding error occurs. That is, when decoding a variable-length code, it is detected that a code with an impossible pattern is received and it is regarded as a decoding error. When such a decoding error occurs, the run level output subsequently decoded is not reliable because the cutout position of the variable length code is incorrect.

Therefore, generally, in an image data decoding device, the following header (in the case of MPEG, a slice header or a sync mark in MPEG4) is used.
r) is searched. Then, based on the next header obtained by the search, the correct decoding process for the variable length code bit string after this header is restored.
However, even if a bit error or the like occurs in the code bit string of the compressed image data, or if the variable length code is not cut out correctly in the variable length code decoding unit, Often any variable length code in the variable length code table is suitable.

When the codes of the variable length code table match in this way, a variable length code decoding error is not detected, and therefore an incorrect run level output is continued. And
Such illegal variable-length code hits continued, and eventually,
Only then will a decoding error be detected. Therefore, the reproduced image data obtained by decoding during the period from the time when the error actually occurred before the detection of the decoding error until the time when the decoding error is detected and the normal run level output is obtained becomes invalid. .

Therefore, when a decoding error is detected at the time of decoding a block in the variable-length code decoding unit, for example, in the motion compensation unit, data having a high probability of being incorrect a few blocks before is used as a reference image for motion compensation. It may have been done. In this way, by performing decoding processing on incorrect data, not only is wasteful processing increased, but the distortion of the decoded image due to decoding errors is larger, and the disturbance continues for a longer period of time. There is also a possibility that quality will be deteriorated.

Therefore, when a decoding error of the variable length code is detected in the variable length code decoding process for the code bit string of the moving image compressed data composed of the variable length code, the predetermined error to which the decoding error occurrence data belongs is detected. This is replaced with the data of the corresponding unit data area of the past decoded reproduction image, and the decoding reproduction image is formed based on the data created by the replacement.

Alternatively, a decoded image obtained by re-decoding from the available information is written to form a decoded reproduced image. This is to correct something that can be corrected by some means even if there is an error so that the deterioration of the image quality of the portion that could not be decoded correctly becomes inconspicuous. MPE
In G4 etc., using the information that could be correctly decoded,
A technique called concealment is used to conceal the part that could not be decoded correctly so that it is not conspicuous. One of them is a tool called data partitioning. The error resilience is improved by performing processing such as changing the transmission order of the information in the video packet, and even when an error occurs, the information such as the particularly important coding mode and motion vector of each macroblock is used to transfer the packet data. It is also possible to perform re-decoding to form a decoded reproduced image. The probability that the first half of the video packet will be correctly decoded is higher than the latter half of the information. Therefore, in data partitioning, of macroblock information included in a video packet, particularly important information is in the first half, and other information is in the second half.

Taking the case of INTRA-VOP as an example, for example, the coding mode, the quantized difference value, and INT required for decoding the AC component in the first half with the INTRA-DC component.
Even if the RADC component is in the first half and the coding pattern, AC prediction flag, AC component, etc. necessary for decoding the AC component cannot be decoded due to an error, the first half of INTRADC
By decoding using only the components, etc., rather than discarding all data of the macroblock and displaying using the decoded image data of the same macroblock number of the previous frame,
The image quality deterioration can be made inconspicuous.

The decoded image data is written in the display memory, the resolution is converted according to the resolution of the display unit, and then the decoded image data is sent to the display unit and displayed as an image. In general, the display means of a mobile information terminal such as a mobile phone has a smaller screen and a lower resolution than a standard TV and the like, so an image suitable for the standard TV cannot be displayed as it is. It is necessary to convert the image to the resolution that matches the copy.

3 and 4 are block diagrams showing the structure of a conventional image compression data decoding apparatus. As shown in FIG. 3, the image data decoding unit 1 uses the reference memory 2 to decode all the compressed image data, read all the decoded image data into the display memory 6, and load the image data in the display memory 6 The resolution-converted image is input to the resolution conversion unit 3, converted into a resolution suitable for the display unit 7, and the image data whose resolution has been converted by the resolution conversion unit 3 is output to the display unit 7. It is displaying.

However, this system has a problem that it is inefficient because it has a display memory for reading all the image data, although it does not use all the information of the decoded image data. Therefore, as shown in FIG. 4, the image data decoding unit 1 uses the reference memory 2 to decode the compressed image data for each block, and the image data decoded for each block is sequentially converted into the resolution conversion unit 3. Is input to the display memory 6 and the resolution-converted image data is output from the display memory 6 to the display unit 7. It is considered.

In the case of the method shown in FIG. 3 in which the resolution of an image that has been entirely restored is converted, data is sequentially transferred line by line from the left end of the restored image, and the pixels to be subjected to resolution conversion are vertically and horizontally. The resolution conversion is easy by performing the process of interpolating the data of 1 to reduce the data. On the other hand, in the method shown in FIG. 4 in which the resolution conversion is performed and then the data is written in the display memory 6, since the processing is performed for each block, the head data of the data packet is not always the leftmost block of the image. , The decoded and resolution-converted data cannot be sequentially stored in the display memory. Further, when a decoding error occurs, only the data in the display memory corresponding to the position where the error occurs cannot be corrected and the image quality deteriorates.

That is, when a decoding error occurs, the packet to which the error block belongs is detected and the data of the packet of the corresponding immediately preceding frame is replaced, or the information available from the data of the obtained packet is used. Then, the packet data is re-decoded, the data is written in the display memory 6, and is displayed on the display unit 7.

At this time, since the head of the packet does not always start from the left end of the image area, the resolution conversion unit 3 which processes in units of lines also has an error in the data of the block from which the normal decoded data is obtained. It is overwritten when the packet data is decrypted again and rewritten.

[0018]

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and reduces the memory capacity by reducing the image by using resolution conversion and then storing the display memory data. When a decoding error occurs, the problem that the image quality is deteriorated because the data in the display memory corresponding to the position where the error occurs cannot be corrected is solved. It is an object of the present invention to provide an image compression data decoding device which facilitates writing to a display memory and can reduce a disorder of a decoded image which occurs when a variable length code decoding error occurs.

[0019]

An image compressed data decoding apparatus according to the present invention includes an image data decoding unit for decoding image compressed data, a display unit for displaying the decoded image data, and a display unit for displaying the decoded image data. Resolution conversion means for converting the resolution of the image data, a display memory for storing the resolution-converted data, a macroblock number output means for outputting a macroblock number indicating the position of the macroblock of the image-compressed data, A write control means for controlling the address of the display memory for writing the decoded and resolution-converted image data based on the output macroblock number, and when a decoding error occurs,
The present invention is characterized in that a decoded image obtained by past decoding from a past decoded reproduced image or available information is written in macroblock units in an area of the display memory corresponding to a macroblock having an error.

By reducing the image using resolution conversion and then storing the display memory data, the memory capacity is reduced, and when a decoding error occurs, write control is performed in macroblock units. As a result, the image data can be written from the middle of the line buffer in the resolution conversion means, and only the data in the display memory corresponding to the position where the error occurs can be corrected, and the disorder of the decoded image can be reduced.

In the image compressed data decoding apparatus of the present invention, when a decoding error occurs, the writing control means writes to the area of the display memory corresponding to the macroblock immediately before the macroblock immediately before the errored macroblock. It is characterized by performing control for masking. By performing the masking control, it is possible to prevent invalid rewriting of the display memory. The image compressed data decoding device of the present invention is characterized in that the operation of converting the resolution and writing it in the display memory is performed only when the image is updated. By reducing the write operation, power consumption can be reduced.

[0022]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. FIG. 1 is a block diagram showing the configuration of an image compression data decoding device according to an embodiment of the present invention. FIG. 1A shows the configuration of the entire image compression data decoding device, and FIG. 1B shows the configuration of the image data decoding unit 11. The image compression data decoding device includes an image data decoding unit 11 that decodes image compression data, a reference memory 12 that is referred to when decoding the image compression data, and a resolution conversion unit 13 that converts the resolution to match the display unit. A macroblock number output unit 14 that outputs a macroblock number corresponding to image data, a display memory write control unit 15 that controls writing of the decoded image data to a display memory based on the macroblock number, and a decoding unit It mainly comprises a display memory 16 for writing the generated image data and a display unit 17 for displaying the image data. Further, the image data decoding unit 11 refers to the variable length code table from the supplied variable length code bit string, cuts out the variable length code, and outputs the quantized DCT coefficient code string, the variable length code decoding processing unit. 18, an inverse quantization unit 19 that inversely quantizes the code string of the quantized DCT coefficient from the variable length code decoding processing unit 18, and the inversely quantized DCT coefficient that is inversely converted into data for each pixel before DCT conversion. An inverse discrete cosine transform unit 20 for transforming, and a motion compensating unit 21 for predicting motion and performing motion compensation for image data obtained by inverse discrete cosine transform.
Mainly consists of and.

The operation of the image compression data decoding apparatus having the above configuration will be described. The code bit string of compressed video data obtained by storage media, communication, etc.
It is supplied to the image data decoding unit 11, and first, the header of each layer is detected from the coded bit string having a hierarchical structure. Here, the hierarchical structure of the image compression data is, for example, MPEG4.
For example, the structure is as shown in FIG. Each picture layer has a minimum unit of 1 with a start code.
It is divided into packets of 6 pixels width and arbitrary length (unit data tilt), and this packet further comprises an arbitrary number of macroblocks (MB), and the macroblock layer further comprises 6 8 × 8 pixel units. It is composed of blocks. That is, four Y signal blocks corresponding to each block space and one Cr signal and one Cb signal used for motion compensation.
It is composed of blocks.

Based on the detected headers of the respective layers, the code bit string is separated for each parameter, and the code bit string arranged in each block with a block of 8 × 8 pixel unit as the minimum decoding processing unit is a variable length code decoding processing unit. 18 are supplied. The variable-length code decoding processing unit 18 cuts out the variable-length code from the supplied variable-length code bit string by referring to the variable-length code table and outputs the corresponding run level output, that is, the quantized DCT coefficient code string. To do. Further, the variable length code decoding processing unit 18 refers to the variable length code table and, if the corresponding variable length code is not entered, generates a decoding error detection signal of the variable length code and sends this to the motion compensation unit 21. Output.

Quantization D from the variable length code decoding processing unit 18
The CT coefficient code string is supplied to the inverse quantization unit 19 and inversely quantized, and the inverse quantized DCT coefficient is inversely transformed by the inverse discrete cosine transform unit 20 into data for each pixel before DCT transformation. . The motion compensation unit 21 performs motion compensation on the image data obtained by the inverse discrete cosine transform, and stores it in the reference memory 12.

In the present invention, the image data decoding unit 11
The display memory write control unit 15 controls writing of the decoded image data to the display memory 16 based on the macro block number from the macro block number output unit 14 which indicates the position of the macro block currently being decoded. In this case, an address is generated so that data can be written from the area of the display memory corresponding to the beginning of the packet in which the decoding error has occurred, in units of macroblocks.

The image data decoding unit 11 decodes the compressed image data for each macroblock, and the decoded image data for each macroblock is sequentially input to the resolution conversion unit 13, and the resolution conversion unit 13 sets the resolution. The converted image data for each macro block is input to the display memory 16, and the resolution-converted image data is displayed from the display memory 16.
Output to 7. At that time, the display memory 16 includes the display unit 17
Since the capacity is reduced according to the resolution of, the output image data output from the image data decoding unit 11 is reduced by the resolution conversion unit 13 and written in the display memory 16.

Therefore, when a decoding error occurs, the packet to which the error-occurring block belongs is detected and the data of the corresponding packet of the immediately preceding frame is replaced, or the information available from the data of the obtained packet is used. And re-decode the packet data to display memory 16
The data is written in and displayed on the display unit 17.

The head of the packet does not always start from the left end of the image area, but since the display memory writing control unit 15 processes macro blocks as a unit,
It is possible to control so that the data of only the macroblock in which the error is detected is re-decoded and rewritten instead of rewriting the entire packet in which the error has occurred.

That is, in units of macroblocks,
Detect the macro block to which the error block belongs,
An address is generated so that data can be written from the area of the display memory corresponding to the beginning of the macroblock in which the decoding error has occurred, and the macroblock data of the corresponding immediately preceding frame is replaced, or the obtained macroblock data It is only necessary to re-decode the macroblock data using the information available from, to write the data to the display memory 16 and display it on the display unit 17.

The processing of the resolution conversion unit 13 is performed on a line-by-line basis, but writing is masked up to the position of the display memory 16 corresponding to the macroblock number immediately before the macroblock in which the error is detected, or By masking the writing up to the position of the display memory 16 corresponding to the immediately preceding packet, invalid rewriting of the display memory 16 can be prevented. Further, comparing the processing of the resolution conversion unit 13 with the conventional method of FIG. 3 in which data is sequentially transferred line by line from the left end of the restored image,
When a decoding error occurs, image data can be written from the middle of the line buffer (a decoded image reproduced in the past or a decoded image obtained by re-decoding from available information).

As described above, in the present embodiment, when a decoding error of a variable length code is detected, the macroblock including the generation area is used as a unit data area and the macroblock including the decoding error is detected. The above unit data area or the data of one or more corresponding unit data areas of the past normal picture is replaced, or one or more of the unit data areas is used by using the available information from the obtained information. It is possible to display the decoded data obtained by re-decoding the data.By controlling the macroblock as a unit area in this way, the decoded display image is affected for a long period of time by the occurrence of decoding error. It is possible to avoid the problem that a large image is disturbed.

Further, the output data of the variable length code decoding unit 18 and the output data of the image data decoding unit 11 are compared with the output data in the previous frame to detect the presence or absence of the image data update for each frame, The power consumption can be reduced by converting the resolution only when the image is updated, that is, by performing the operation of converting the resolution from the reference memory and writing to the display memory only when the image is updated. Further, if the operation of detecting the presence or absence of image data update in macroblock units and converting the resolution and writing to the display memory is performed only when the image is updated in macroblock units, the image changes within the frame. However, since there may be a macroblock in which the image does not change, it is possible to further reduce power consumption.

[0034]

As described above, according to the image compression data decoding apparatus of the present invention, the image capacity is reduced by using the resolution conversion, and then the display memory data is accumulated to reduce the memory capacity. When a decoding error occurs, the problem that the image quality is deteriorated because the data in the display memory corresponding to the position where the error occurs cannot be corrected is solved. The writing to the display memory can be facilitated, and the disorder of the decoded image generated at the time of the decoding error of the variable length code can be reduced.

According to the image compression data decoding apparatus of the present invention, even when the decoded data is reduced by the resolution conversion unit and the display data is written in the display memory, when the decoding error of the variable length code is detected. By replacing it with the data of the macroblock containing the generation area or the data of the corresponding macroblock of a normal picture in the past, or by re-decoding using the available information from the obtained information. It is possible to display the obtained decoded data, and it is possible to avoid a problem that a decoded display image is affected for a long period of time due to the occurrence of a decoding error, or a large image is disturbed. Also,
Since the display data is written to the display memory after the resolution conversion, the memory capacity of the display memory can be reduced without converting the resolution each time the display data is transferred to the display unit.

Further, when a decoding error occurs, image data can be written from the middle of the line buffer in units of macroblocks, and the display memory corresponding to the block before the erroneous packet in the display memory can be written. Since the invalid data is not overwritten, it is possible to make the disorder of the decoded image inconspicuous in the decoding error of the variable length code with the memory capacity as small as possible. Moreover, the power consumption can be reduced by converting the resolution only when the frame is updated.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an image compressed data decoding device according to an embodiment of the present invention.

FIG. 2 is a schematic diagram for explaining a hierarchical structure of image data in the MPEG system.

FIG. 3 is a block diagram showing a configuration of a conventional image compressed data decoding device.

FIG. 4 is a block diagram showing a configuration of a conventional image compression data decoding device.

[Explanation of symbols]

1 Image data decoding unit 2 Reference memory 3 Resolution converter 6 Display memory 7 Display 11 Image data decoding unit 12 Reference memory 13 Resolution converter 14 Macroblock number output section 15 Display memory write controller 16 display memory 17 Display 18 Variable Length Code Decoding Processor 19 Dequantizer 20 Inverse Discrete Cosine Transform Unit 21 Motion Compensation Unit

   ─────────────────────────────────────────────────── ─── Continued front page    F-term (reference) 5C059 KK01 KK08 KK38 KK49 LA06                       LB05 MA05 MA23 MC11 ME01                       NN01 RB02 RC22 RF01 RF09                       UA05 UA37 UA38                 5C076 AA22 BA03 BA06 CB04                 5J064 AA01 BA09 BB08 BC01 BC16                       BD03

Claims (3)

[Claims] 1. An image data decoding means for decoding image-compressed data, a display means for displaying the decoded image data, a resolution conversion means for converting the resolution of the image data according to the display means, and a resolution. A display memory for storing the converted data, a macroblock number output means for outputting a macroblock number indicating the position of the macroblock of the image-compressed data, and a decoding and resolution based on the output macroblock number. Write control means for controlling the address of the display memory for writing the converted image data, and in the case where a decoding error occurs, a macroblock unit for the area of the display memory corresponding to the macroblock in error In addition, it is possible to write the decoded image obtained by re-decoding from the past decoded playback image or available information. A characteristic image compression data decoding device. 2. The write control means, when a decoding error occurs, controls the write masking to the area of the display memory corresponding to the macroblock immediately before the macroblock having the error. The moving image compressed data decoding device according to claim 1. 3. The moving image compressed data decoding apparatus according to claim 1, wherein the operation of converting the resolution and writing the same in the display memory is performed only when the image is updated.