JPH10271516A - Compression coder, coding method, decoder and decoding method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the storage capacity of a memory provided in the compression coding/decoding device in an image information transmission system in compliance with the Motion Picture Experts Group(MPEG) or the Joint Photograph Experts Group(JPEG) or the like. SOLUTION: In an encoder 1 as a compression coder applying compression coding to a received image signal in compliance with the standards such as the MPEG, other compression/expansion is conducted by using a signal compression circuit 11 and a signal expansion circuit 12 than the main compression coding which is conducted by using a motion detection/compensation processing circuit 9, a discrete cosine transformation/quantization circuit 4, a Huffman coding circuit 5. Thus, the storage capacity of the memory 2 is reduced by reducing the information quantity to be written in a memory 2 provided to the compression coder. Moreover, the storage capacity of the memory 2 provided to the decoder by applying other compression/expansion to the decoder in compliance with the MPEG or the like other than the decoding processing corresponding to the compression coding is reduced.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention
The present invention relates to a compression encoding device and an encoding method for compressing an image signal with high efficiency, and a decoding device and a decoding method for decoding image information in a system for transmitting image information conforming to the regulations of G, JPEG and the like.

[0002]

2. Description of the Related Art In a system for transmitting image information, a compression encoding device for compressing and encoding an image signal, and a decoding device for decoding an image signal from a compression-encoded signal are indispensable elements. In order for the compression encoding device and the decoding device to support multimedia, that is, to be applicable to various image information transmitting means or receiving means, MPEG, JPEG,
Etc. are defined.

[0003] MPEG is a regulation on a method of transmitting a moving image such as a video image. As a compression encoding method,
A motion compensation inter-frame predictive coding method for performing information compression using temporal correlation, and a discrete cosine transform (hereinafter, DCT) as orthogonal coding for performing information compression using spatial correlation.
) And a Huffman coding method as entropy coding (variable-length coding) that performs information compression using bias in the appearance probability of codes generated by the above-described coding method. It is a thing.

[0004] Among them, the motion-compensated inter-frame predictive encoding method is required in a system for transmitting moving images, and is adopted in MPEG and the like. The motion-compensated inter-frame predictive coding is performed by referring to other frames in units of each frame in the signal. MPEG
In, regarding motion-compensated inter-frame prediction coding,
For each frame, I picture, P picture and B picture
Three picture types of pictures are defined.

[0005] An I picture is an image that is intra-coded, that is, encoded without referring to other frames.
It is not subject to motion compensated inter-frame prediction coding. A P picture is an image that is subjected to forward prediction, that is, motion-compensated inter-frame prediction coding by referring to a preceding frame. Further, a B picture is an image for which bidirectional prediction is performed. That is, this is an image on which motion-compensated inter-frame prediction coding is performed by referring to the preceding frame and the following frame.

[0006] The picture types of frames referred to for motion-compensated inter-frame predictive coding of other frames are I-pictures and P-pictures. B pictures cannot be referenced. In other words, a frame referred to for performing motion compensation inter-frame predictive coding on a P picture is a preceding I picture or P picture. Further, in order to perform motion compensation inter-frame predictive coding on a B picture, it is necessary to refer to a preceding frame (I picture or P picture) and a subsequent frame (I picture or P picture).

Accordingly, the process of motion-compensated inter-frame predictive coding of a B picture is performed in a frame (I picture or P picture) at a subsequent position in an input image signal.
Can only be performed after the has been processed. For this reason, the order of each frame in the input image signal is different from the processing order. Then, after the order of the frames in the input image signal is rearranged (reordering), the frames are supplied to a means for performing motion compensation inter-frame predictive coding. As will be described later, a memory is used when performing such rearrangement.

[0008] A compression encoder according to the MPEG standard
After performing such motion compensation inter-frame prediction coding, compression coding is performed by performing coding by DCT as described above and Huffman coding as entropy coding (variable length coding). Generate a signal. Then, such a compression-coded signal is transmitted.

Therefore, an image signal generated by decoding the compressed and coded signal transmitted by the decoding device is obtained by rearranging the order of frames.
For this reason, a process of restoring the order of the frames is required. As will be described later, when performing such processing to restore the order of the frames, a memory provided in association with the decoding device is used.

[0010] On the other hand, JPEG is a regulation concerning still image transmission such as facsimile transmission. Unlike MPEG which handles moving images, there is no need to perform a motion compensation inter-frame predictive encoding method.
It is assumed that the encoding method using T and the Huffman encoding method are combined.

The processing by the compression encoding / decoding device as described above proceeds by transferring data to and from a memory as shown in FIG. Therefore, a memory is provided in association with the compression encoding / decoding device. The role played by such a memory will be described below.

First, as described above, in a compression encoding apparatus that handles moving images, input is performed in order to rearrange the order of frames required due to the use of the motion-compensated inter-frame prediction encoding method. Such a memory is used for temporarily storing an image signal. That is, control is performed so that the order of taking out from the memory is suitable for performing motion compensation inter-frame predictive coding. In addition, such a memory also serves to store a frame (a local decoded image to be described later) which is referred to as a reference for performing the motion-compensated inter-frame predictive encoding method.

Further, such a memory is used for recognizing the properties of each frame, for example, including a scene change.
Prior to processing by the compression encoding / decoding device, it is also used to secure an image signal for a predetermined time. As described above, the use of the memory is performed in a compression encoding device that handles both moving images and still images.

On the other hand, a memory provided in association with a decoding apparatus for handling moving images firstly stores, in the decoding corresponding to the compression encoding as described above, the decoding corresponding to the motion compensation inter-frame prediction encoding. Used to store the signal referenced as a reference. It is also used to temporarily store the decoded image signal for the process of returning the frame order to the original image signal after decoding is completed. That is, control is performed so that the order of extracting the decoded image signal from the memory is the order of the original image signal.

A memory provided in association with such a compression encoding / decoding device stores several frames to several tens of frames depending on conditions such as the performance of an image transmission system and the properties of an image to be handled. Is done. For this reason, in general,
As such a memory, a memory having a large storage capacity is required. For example, a compression encoding / decoding device conventionally used is at least 16 Mbit to 32 Mbit.
In the case of many bits, a memory of 320 Mbits is required.

[0016]

SUMMARY OF THE INVENTION As described above, MP
A compression encoding / decoding device used in an image information transmission system conforming to the regulations of EG, JPEG and the like requires a memory having a large storage capacity for performing processing. For this reason, the cost of the memory provided in association with the compression encoding / decoding device has been one of the factors that increase the cost of the entire system.

Accordingly, an object of the present invention is to provide MPEG, J
In an image information transmission system according to the provisions of PEG, etc.,
A compression encoding device capable of reducing the capacity of a memory provided in association with the compression encoding device and the decoding device,
An object of the present invention is to provide an encoding method, a decoding device, and a decoding method.

[0018]

According to a first aspect of the present invention, there is provided a compression encoding apparatus for encoding an image signal, comprising: a memory;
A compression circuit for compressing the input image signal, an expansion circuit for expanding the signal compressed by the compression circuit, and writing of the signal compressed by the compression circuit to a memory;
And a memory control circuit for controlling reading and reading, and a coding circuit for coding an image signal expanded by the expansion circuit.

According to a twelfth aspect of the present invention, there is provided a decoding apparatus for decoding an encoded image signal, comprising: a memory; a decoding circuit for decoding the encoded image signal; A compression circuit for compressing an image signal decoded by the circuit, an expansion circuit for expanding the signal compressed by the compression circuit, and a memory for controlling writing and reading of the signal compressed by the compression circuit with respect to the memory And a control circuit.

According to a seventeenth aspect of the present invention, in the encoding method for encoding an image signal, a compression step of compressing the input image signal by a predetermined compression method, and a step of writing the compressed signal to a predetermined memory. Reading a compressed signal stored in a predetermined memory, decompressing the read signal, and encoding the transmitted image signal for transmission. An encoding method characterized by the following.

According to a twenty-eighth aspect of the present invention, in the decoding method for decoding an encoded image signal, a decompression decoding step of decoding the encoded image signal and a decoding step of storing the decoded image signal in a predetermined memory. Compressing the compressed image signal by a predetermined compression method, writing the compressed signal to a predetermined memory, reading a compressed signal stored in the predetermined memory, And a decompression step for decompressing the signal.

According to the first and seventeenth aspects of the present invention, the amount of information of a signal to be written into a memory provided in association with a compression encoding device is determined by the main compression encoding performed according to the MPEG standard or the like. It can be reduced by performing another compression / decompression. Therefore, the required memory capacity can be reduced.

According to the twelfth and twenty-eighth aspects of the present invention, the information amount of a signal written in a memory provided in association with the decoding device corresponds to the main compression encoding performed according to the MPEG standard or the like. This can be reduced by performing compression / decompression separately from decoding. Therefore, the required memory capacity can be reduced.

[0024]

BEST MODE FOR CARRYING OUT THE INVENTION

[First Embodiment] A first embodiment of the present invention will be described below. The first embodiment of the present invention is an application of the present invention to a compression encoding device in a transmission system conforming to the MPEG regulations. FIG. 1 shows a block diagram of a first embodiment of the present invention. A memory 2 is provided in association with the encoder 1 as a compression encoding device.

The pre-processing unit 3 receives an image signal and performs a predetermined process on the received image signal. The motion detection / compensation processing circuit 9, the DCT / quantization circuit 4, and the Huffman coding circuit 5 perform a compression coding process described later on the output of the preprocessing unit 3. The inverse DCT / inverse quantization circuit 10 performs processing opposite to that of the DCT / quantization circuit 4 in order to generate a signal referred to for processing by the motion detection / compensation processing circuit 9 as described later. . Further, the memory controller 6 includes each component in the encoder 1 as described above,
The signal transfer to and from the memory 2 is controlled. Further, the code interface 7 outputs the coded signal as an MPEG video bit stream. Controller 8
Controls the overall operation of the encoder 1.

The signal compression circuit 11 performs compression so that the capacity of the memory 2 can be reduced. Such compression is different from main compression, that is, compression encoding performed according to the regulations of MPEG and the like. Also, the signal expansion circuit 1
2 restores the signal compressed by the signal compression circuit 11 to the original signal.

An operation for processing performed by the first embodiment of the present invention will be described. The input image signal is, for example, converted from (4: 2: 2) to (4: 2:
After a predetermined process such as conversion to 0) is performed, the signal is supplied to the signal compression circuit 11. The signal compression circuit 11 compresses the signal as described later and supplies the signal to the memory controller 6. The memory controller 6 writes the supplied signal to the memory 2. As a result of this writing, a sufficient amount of compressed image signal is stored in the memory 2 to start encoding. Further, as described above, recognition of the property of the image stored in the memory 2 such as including a scene change is performed.

After the conditions for starting encoding have been established in this way, the compressed image signal, that is, the image signal to be subjected to main compression encoding, and the movement of the image signal are detected in accordance with an instruction from the controller 8. For reference, an image signal to be referred to is supplied from the memory 2 to the signal decompression circuit 12. The signal expansion circuit 12 restores the supplied signal, that is, the compressed image signal to the original data, as described later. Then, the restored signal is supplied to the motion detection / compensation processing circuit 9. When the restored image signal is supplied to the motion detection / compensation processing circuit 9 in this manner,
As described above, the order of each frame in the input image signal is rearranged (reordering).

The motion detection / compensation processing circuit 9 performs a motion detection / compensation process (ME / MC) as the above-described motion compensation inter-frame prediction coding method. That is, the signal expansion circuit 1
Motion detection between the signal supplied from the second unit (the image signal to be subjected to main compression encoding and the reference image signal), and the difference between the detected signal and the frame referenced as a standard is calculated in accordance with the detection result. By performing the processing, information compression using the temporal correlation of the supplied signals is performed. This motion detection / compensation processing circuit 9 is supplied with information on image signals from the memory controller 6 as shown by a dotted line as necessary. Thus, the signal encoded by the motion detection / compensation processing circuit 9 is supplied to the DCT / quantization circuit 4. The image signal of the frame referred to in the motion detection is an inverse DCT / inverse quantization circuit 10.
Is held. Note that an image signal requiring intra-frame encoding is supplied to the DCT / quantization circuit 4 as it is.

The DCT / quantization circuit 4 performs DCT as an orthogonal transform on the supplied signal, and further quantizes a pixel value calculated by the DCT. In this way, information compression using the spatial correlation of the supplied signals is performed. The signal encoded by the DCT / quantization circuit 4 is supplied to a Huffman encoding circuit 5 and, at the same time,
It is also supplied to the T / inverse quantization circuit 10.

The Huffman coding circuit 5 performs, for example, two-dimensional Huffman coding on the supplied signal. Huffman coding is a coding method that embodies entropy coding (variable length coding). That is, for a code value generated by the above-described coding with information compression performed by the motion detection / compensation processing circuit 9 and the DCT / quantization circuit 4, a longer code length is set to a value having a lower appearance probability. This is an encoding method in which the amount of information of the entire signal is reduced by assigning and assigning a short code length to a value having a high appearance probability.

The signal generated by the Huffman encoding circuit 5 is
The data is written to the memory 2 via the memory controller 6.
Then, it is output as an MPEG video bit stream at a predetermined transmission timing via the code interface 7. This MPEG video bit stream is the final output of the encoder 1.

On the other hand, as described above, the signal encoded by the DCT / quantization circuit 4 is also supplied to the inverse DCT / inverse quantization circuit 10. Inverse DCT / inverse quantization circuit 10
Performs inverse DCT / inverse quantization on the supplied signal. From the signal subjected to the inverse quantization and the image signal supplied from the motion detection / compensation processing circuit 9, a local decoded image referred to in the processing performed by the motion detection / compensation processing circuit 9 is generated as described later. You. As described below,
The output of the inverse DCT / inverse quantization circuit 10 is selectively written into the memory 2 as a local decoded image. No compression is performed during such writing. Note that the image signal that has been subjected to intra-frame encoding is written as it is as a memory.

The above-described writing to the memory 2 will be described with reference to FIG. As described above, the preprocessing unit 3
Supplies an image signal to the signal compression circuit 11. At this time, irrespective of the picture type, all the frames in the input image signal are supplied to the signal compression circuit 11 and written into the memory 2 after being compressed. In the subsequent stage, that is, motion detection /
Compensation processing circuit 9, DCT / quantization circuit 4, and inverse DCT
The processing by the inverse quantization circuit 10 is generally indicated as 30 in FIG. By this processing, a locally decoded image is generated based on the I picture and the P picture, and written into the memory 2. That is, for the I picture, the processing result of the above-described processing 30 is written to the memory 2 as it is. Also, for P pictures,
The result of the processing 30 and the local decoded image signal read from the memory 2 are added to each other and then written into the memory 2.

As described above, there are three picture types defined in MPEG or the like: I picture, P picture, and B picture.
As described above, a B picture is encoded with reference to another frame. That is, with respect to the P picture and the B picture, the motion detection / compensation processing circuit 9 calculates a difference from the above-described local decoded image, and performs motion compensation inter-frame predictive coding based on the calculated difference. Therefore, the locally decoded image is supplied to the motion detection / compensation processing circuit 9 as shown in FIG. As described above, since the locally decoded image is not compressed when it is written to the memory 2, it is not expanded when it is supplied from the memory 2 to the motion detection / compensation processing circuit 9.

On the other hand, in general, when a certain signal is compressed / expanded, it does not completely return to its original state, but contains some error. As far as the transmission of individual frames is concerned, such an error does not pose any problem as long as it is within an allowable range for the image quality required of the transmission system. However, as described above, since the locally decoded image is involved in encoding other frames, the local decoded image is compressed /
Once decompressed, the error caused by compression / decompression is
It can affect the coding of a wide portion of the image signal and can be a factor in reducing the overall quality of the transmitted image. For this reason, as described above, when the I picture and the P picture used as the local decoded image are written to the memory 2, and when the motion detection /
When supplied to the compensation processing circuit 9, compression / expansion is not performed.

The compression other than the main compression, that is, the compression performed by the signal compression circuit 11 will be described in detail.
Such compression is performed by, for example, a differential pulse code modulation method (hereinafter referred to as
DPCM). In the first embodiment of the present invention, 40 bits corresponding to 5 pixels (that is, 8 (bits / pixel) × 5 pixels = 40 bits) are compressed to 32 bits. FIG. 3 shows a signal compression circuit 11.
The detailed configuration of is shown. The switching control circuit 110 switches the supply destination for each pixel by controlling the switch 44, the switch 45, and the multiplexer 48.

The image signal is supplied for each pixel, that is, in units of 8 bits. When the first pixel is supplied, the switch 4 is turned on by a command from the switching control circuit 110.
4 connects the terminal 49 and the terminal 51. Therefore, the supplied 8-bit pixel data is supplied to the multiplexer 48 and the switch 45. Further, when the first pixel data is supplied, the switch 45 connects the terminal 52 and the terminal 54 according to a command from the switching control circuit 110. Therefore, the first pixel data is supplied to the flip-flop 42.

When the second pixel is supplied, the second pixel data and the flip-flop 42 are supplied as described above.
Is subtracted from the first pixel data stored in the
6 is calculated. This difference is supplied to the quantizer 41 and is quantized to 6 bits. By the way, when the second pixel is supplied, the switch 44 connects the terminal 50 and the terminal 51 according to a command from the switching control circuit 110. For this reason, the 6-bit data generated as described above is supplied to the inverse quantizer 43 while being supplied to the multiplexer 13.

The inverse quantizer 43 inversely quantizes the supplied 6-bit data and supplies it to the adder 47. Adder 4
Reference numeral 7 adds the data supplied from the inverse quantizer 43 and the first pixel data stored in the flip-flop 42 to restore the second pixel data. By the way, when the second pixel is supplied, the switch 45 connects the terminal 53 and the terminal 54 according to a command from the switching control circuit 110. Therefore, the result of the above addition, that is, the restored second pixel data is stored in the flip-flop 42. In this way,
The restored pixel data stored in the flip-flop 42 is used in the subtractor 46 and the adder 47 when processing the next input pixel data.

Further, the third to fifth pixel data are processed in the same manner as the second pixel data, and each data is made up of 6 bits. As a result, as shown in the lower part of FIG. 3, the data (40 bits) for five pixels is compressed to 32-bit data. Each time 32-bit data is generated in this way, it is output to the memory 2 via a bus capable of transferring 32 bits. Thus, the compressed data corresponding to five pixels is written to one address on the memory 2.

Thereafter, the sixth and subsequent pixel data are compressed into 32-bit data in the same manner as the above-described operation, using 40-bit data of five pixels as a processing unit. Therefore, the image signal input to the signal compression circuit 11 is always compressed by 4/5.

Next, from the data compressed by the signal compression circuit 11, which is different from the main compression,
Decompression for restoring original data will be described. Such expansion is performed by the signal expansion circuit 12. The basic configuration is to perform reverse processing of the compression performed by the signal compression circuit 11. Therefore, in the first embodiment of the present invention, the process of returning the 32-bit data generated as described above to the original 40-bit data is performed. FIG. 4 shows a detailed configuration of the signal expansion circuit 12.
The switching control circuit 120 controls the switching of the supply destination on a pixel-by-pixel basis by the demultiplexer 61 and the switch 64.

As described above, the 32-bit data fetched from the memory 2 according to a command from the controller 8 is supplied to the demultiplexer 61. Demultiplexer 6
1 is 8-bit pixel data of the supplied 32-bit data (in the above-described compression, the pixel data whose bit length remains 8 bits, that is, the first pixel data of 5 pixels which is a processing unit) (Pixel data) is processed first, and then the pixel data is supplied to the subsequent stage so that 6-bit data is processed in order from the second.

When 8-bit data is output from the demultiplexer 61 as the first data, the switch 64 is switched to the terminal 6 by an instruction from the switching control circuit 120.
6 and the terminal 68 are connected. Therefore, the 8-bit data passes through the signal expansion circuit 12 and is stored in the flip-flop 65.

When 6-bit data is output from the demultiplexer 61 as the second data,
The 6-bit data is inversely quantized by the inverse quantizer 62 and supplied to the adder 63. The adder 63 is connected to the data supplied from the inverse quantizer 62 and the flip-flop 65
, And the second data is restored. By the way, when 6-bit data is output as the second data from the demultiplexer 61, the switch 64 connects the terminal 67 and the terminal 68 according to a command from the switching control circuit 120. For this reason,
The output of the adder 63, that is, the restored second pixel data passes through the signal expansion circuit 12 and is stored in the flip-flop 65. In this manner, the restored pixel data stored in the flip-flop 65 is used in the adder 63 when processing the next input pixel data.

Thereafter, the same processing is performed on the third to fifth data. As a result, as shown in the lower part of FIG. 4, 32-bit data is expanded to 40-bit data, and is returned to data of 5 pixels of 8 bits / pixel. Thereafter, the data supplied to the signal decompression circuit 12 is also returned to 40 bits, that is, data for 5 pixels, by combining the 32-bit data into one unit in the same manner as the above operation. Therefore, the signal expansion circuit 1
2 is always expanded by 5/4.

In the first embodiment of the present invention, 4/5 DPCM compression for compressing 40 bits of 8 pixels into 32 bits is performed as compression / expansion different from main compression / encoding. However, in addition to the above, it is possible to use an optimal compression ratio from the viewpoints of a memory capacity to be reduced, a quality required for a transmitted image, and the like. The present invention is also applicable to a case where an image signal having a data length other than 8 bits per pixel is handled.

The above-described compression / expansion can be performed on both the luminance signal and the chroma signal, or can be performed only on the chroma signal. Which method is to be applied may be determined according to the amount of memory to be reduced, the quality required for the transmitted image, and the like. Further, a compression method other than the DPCM compression may be used as long as it is effective in consideration of the amount of memory to be reduced and the quality required for the transmitted image.

FIG. 5 shows the amount of memory used in the first embodiment of the present invention. FIG. 6 shows a case where the present invention is not applied. Here, the effect of applying the present invention to a compression encoding device configured to store five frames of original images prior to compression encoding will be described.

In the case where the present invention shown in FIG. 6 is not applied, the ratio between the original image and the other parts (local decoded image and others) is 65:41. On the other hand, when the compression of 4/5 is performed only on the chroma signal in the first embodiment shown in FIG. 5, the information amount of the original image is compressed. Ratio 59:
41. Such compression of the information amount of the original image can be achieved by compression encoding different from the main compression as described above. 5 and 6 show the memory capacity used in each case. According to these figures, the memory capacity actually reduced is 4 Mbit.

In the case of FIG. 6, one 4 Mbit memory is required in addition to the two 16 Mbit memories. That is, in this case, three memories are required. On the other hand, in FIG. 5, since the required memory capacity is reduced, it is possible to perform compression encoding by using two 16 Mbit memories. For this reason, in FIG. 5, when the compression encoding device is designed as one chip, the number of pins and buses of the LSI for transferring data between the memory and the memory is also reduced. That is, by applying the present invention, the effect that the overall configuration of the compression encoding device is simplified can be obtained.

[Second Embodiment] In the first embodiment of the present invention as described above, as described with reference to FIG. 2, the input image signal is separated from the main compression. In other words, when the data is compressed by the signal compression circuit 11 and written into the memory, the compression is performed on all portions in the image signal regardless of the picture type. On the other hand, as a second embodiment of the present invention, it is possible to selectively perform such compression according to the picture type.

The overall configuration of the second embodiment of the present invention is the same as that of the first embodiment of the present invention described above with reference to FIG. However, the information on the picture type is supplied from the preprocessing section 3 to the memory controller 6 and further to the motion detection / compensation processing circuit 9 via the path shown by the dotted line in FIG. The handling of image signals in the signal compression circuit 11 and the signal expansion circuit 12 will be described with reference to FIG. FIG. 2 used in the description of the first embodiment of the present invention.
Similarly to FIG. 7, FIG. 7 shows the processing by the pre-processing unit 3, and the processing 30 by which the processing by the motion detection / compensation processing circuit 9, the DCT / quantization circuit 4 and the inverse DCT / inverse quantization circuit 10 in the subsequent stage is generally shown. Is shown. As shown in FIG.
In the second embodiment of the present invention, compression / expansion other than the main compression is applied only to B pictures, but not to I and P pictures.

That is, in FIG. 7, only the B picture is supplied from the preprocessing unit 3 to the signal compression circuit 11 and is written into the memory 2 after being compressed. The I picture and the P picture are written to the memory 2 without being compressed. In the subsequent process, as described above, the I picture and the P picture are read from the memory 2 and are not decompressed, but are directly subjected to the process generally shown as 30 as described above. . However, the P picture is processed after the local decoded image is subtracted. The B picture is read from the memory 2 and decompressed before being supplied to the signal decompression circuit 12. After the local decoded image is subtracted from the B picture expanded by the signal expansion circuit 12, the image is processed by the above-described processing 30.

The processing result based on the I picture in the processing 30 is written in the memory 2 as it is. Further, the local decoded image read from the memory 2 is added to the processing result based on the P picture by the processing 30 and then written into the memory 2. The processing result of the processing 30 based on the P picture and the B picture written in the memory 2 in this manner is used as a locally decoded image.

As described above, a B picture is not referred to for handling other pictures in compression encoding / decoding. For this reason, in the main compression, the compression rate of the B picture is generally set to a higher value than the I picture and the P picture. Therefore, the image quality degradation caused by the main compression for the B picture is
It is larger than the picture quality degradation that occurs for pictures and P pictures. In consideration of such a situation, an allowable range of image quality deterioration caused in a B picture due to compression / expansion different from main compression is wider than that of an I picture and a P picture. Therefore, since the I-picture and the P-picture are not compressed / expanded separately from the main compression, the B-picture is compressed / decompressed differently from the main compression.
It is possible to increase the compression ratio of expansion.

In the above-described second embodiment of the present invention, compression / expansion is not performed on I-pictures and P-pictures in consideration of the importance of the locally decoded picture. By the way, depending on conditions such as quality required for a transmission image, a certain error may be allowed even in consideration of being referred to as a local decoded image.
In such a case, it is also possible to perform compression other than the main compression on the I picture and the P picture at a compression ratio such that the error falls within the allowable range.

[Third Embodiment] A third embodiment of the present invention further improves the access efficiency in the main compression. Prior to the description of the third embodiment of the present invention, the relationship between the access efficiency in main compression and the size of a unit for performing compression different from main compression will be described. As described above, DPCM compression is used for compression different from main compression.

In general DPCM compression, compression is performed in units of one frame or one line. This is shown in FIG. In FIG. 8, each horizontal line indicates one line. Further, in FIG. 8, one frame is shown using 15 horizontal lines for simplification. The position marked with a circle is the reset position. That is, only the pixel data at the reset position is transmitted without being compressed, and the difference from the previous data is transmitted for all subsequent pixel data. Therefore, when it is desired to take out only the last pixel data from the compressed data for one frame, for example, it is necessary to sequentially expand the pixel data at the reset position indicated by a circle starting from the start point.

On the other hand, in the signal compression circuit 11 of the first embodiment of the present invention, as shown in FIG. 9, a reset position is set for each unit (5 pixels) smaller than one line. Therefore, when the signal is expanded by the signal expansion circuit 12, it is easy to access data at an arbitrary position.

From the above viewpoint, a third embodiment of the present invention will be described. The overall configuration is the same as that of the first embodiment of the present invention described above with reference to FIG. And FIG.
The signal compression circuit 11 in the middle performs DPCM compression as a compression different from the main compression in a method as described below, so that the DCT / quantization circuit 4 and the inverse DCT / inverse quantization circuit 10 When processing is performed, access to a block at an arbitrary position becomes easy.

The above-described DPCM compression will be described with reference to FIG. If a unit accessed in the main compression (a macroblock in MPEG or the like) is referred to as one block, the above-described DPCM compression reset is performed in a predetermined unit of one block or less. In the following description, such a unit of DPCM compression is referred to as a DPCM block. FIG.
In the diagram shown on the right side in FIG. 0, the position marked with a circle is the reset position. The pixel at the reset position is transmitted with the same data length. In addition, in the diagram shown on the right side in FIG. 10, one DPCM block is composed of three pieces of the line divided into a short line, starting from the reset position.

Therefore, in the example of FIG. 10, in order to access data for one block, it is sufficient to access several DPCM blocks that differ depending on the position of the block. For example, to access block A in the figure, it is sufficient to access four DPCM blocks. The number of DPCM blocks necessary to access one block of data is maximized, for example, when accessing block B. At this time, as shown in the figure, only six DPCM blocks are accessed. Good.

As described above, the size of the DPCM block is set to be equal to or smaller than the size of the main compression unit (the above-described block), and the DPCM block as a different compression from the main compression is set.
By performing the compression, it is possible to perform a different compression from the main compression without reducing the access efficiency in the main compression. Therefore, it is possible to reduce the capacity of the memory provided with the compression encoding device while maintaining the original access efficiency of the image information transmission system to which the present invention is applied.

[Fourth Embodiment] The first embodiment of the present invention described above.
In the embodiment and the like, the feedback processing is performed in the main compression. That is, the main compression is performed by the motion detection / compensation processing circuit 9 and the DCT / quantization circuit 4.
The output of the DCT / quantization circuit 4 is also supplied to an inverse DCT / inverse quantization circuit 10 for inverse DCT / inverse quantization. The I picture and the P picture generated by the inverse DCT / inverse quantization are written in the memory 2 without being compressed, and the motion detection / compensation processing circuit 9 outputs the P picture and the P picture belonging to another frame in the image signal. When a B picture is processed, it is referred to as a locally decoded image.

Therefore, when performing a motion detection / compensation process, which is one step of the main compression, on a series of image signals, the result of the process (inverse DCT / inverse quantum Output of the conversion circuit 10) is involved in the processing of another frame. Therefore, in the above-described first embodiment of the present invention, as shown in FIG. 11, compression / expansion different from the main compression is selectively performed for each frame in the signal. That is, as described above, a signal portion referred to in main processing such as a locally decoded image is not subjected to compression / expansion different from main compression.

On the other hand, as a fourth embodiment of the present invention, the present invention can be applied to a compression encoding device in which feedback processing is not performed in main compression. Such a compression encoder is a JPEG
And the like, which is used in an image transmission system according to the regulations for transmitting still images. Such a system is different from the one that complies with the rules such as MPEG that handles moving images.
There is no need to perform motion compensation inter-frame prediction coding. For this reason, there is no need to use the locally decoded image as described above, and there is no need to perform control such as not performing compression / expansion other than main compression on a specific picture type. Therefore, as shown in FIG. 12, the signal compression circuit 11 and the signal expansion circuit 12 compress / expand all supplied image signals.

FIG. 1 shows the overall configuration of the fourth embodiment of the present invention.
3 is shown. 13, the same components as those in the first embodiment of the present invention described with reference to FIG. 1 are denoted by the same reference numerals. As described above, since it is not necessary to perform motion compensation inter-frame prediction coding, FIG. 13 does not include the motion detection / compensation processing circuit 9 in FIG. Further, since it is not necessary to use the above-described local decoded image, the inverse DCT / inverse quantization circuit 4 shown in FIG.
Are not included in FIG. Further, for the reasons described above, the signal compression circuit 11 and the signal decompression circuit 12
Compress / decompress all the supplied signals.

In this case, all the supplied signals are compressed / expanded at a constant compression ratio, and the memory capacity can be reduced correspondingly.

[Fifth Embodiment] The first embodiment of the present invention described above.
The fourth to fourth embodiments all apply the present invention to a compression encoding device. On the other hand, a fifth embodiment of the present invention described below applies the present invention to a decoding device in a transmission system conforming to the MPEG standard, including the above-described first embodiment of the present invention, for example. It was done. In this way, the capacity of the memory provided in association with the decoding device can be reduced. FIG.
FIG. 11 shows an overall configuration of a fifth embodiment of the present invention. A memory 82 is provided in association with the decoder 81 as a decoding device.

The code interface 87 receives the compressed and coded signal, performs predetermined processing on the received compressed and coded signal image signal, and supplies it to the subsequent stage. Huffman decoding circuit 8
5. The inverse DCT / inverse quantization circuit 84 and the motion compensation processing circuit 89 perform a decoding process as described below according to the rules of MPEG. The motion compensation processing circuit 89 supplies a predetermined result based on the I picture and the B picture among the processing results to the memory controller 86 as a locally decoded image. In this way, the locally decoded image is written into the memory 82, and is read and referred to as needed in the operation of the motion compensation processing circuit 89. The memory controller 86 controls the transfer of signals between the components in the decoder 81 and the memory 82 as described above. The post-processing unit 83 receives the decoded image signal, performs predetermined processing, and outputs the processed signal to the subsequent stage as an output of the decoder 81. The controller 88 controls the overall operation of the decoder 81 as a whole.

The signal compression circuit 91 performs compression so that the capacity of the memory 82 can be reduced. Signal expansion circuit 90
Recovers the original signal from the signal compressed by the signal compression circuit 91.

Next, the operation of the fifth embodiment of the present invention will be described. The transmitted compression-encoded signal, which is input to the decoder 81, is transmitted to the code interface 87.
Is written to the memory 82 via Then, it is extracted in accordance with a command from the controller 88 and supplied to the Huffman decoding circuit 85. The Huffman decoding circuit 85 performs decoding (variable length decoding) corresponding to Huffman coding on the supplied compressed and coded signal.

The signal obtained as a result of the Huffman decoding is supplied to an inverse DCT / inverse quantization circuit 84. Inverse DC
The T / inverse quantization circuit 84 applies the above-mentioned D to the supplied signal.
Decoding by inverse DCT / inverse quantization corresponding to encoding by CT / quantization is performed. Then, the signal decoded by the inverse DCT / inverse quantization is converted to a motion compensation processing circuit 89.
To supply. Of these signals, the I picture and the P picture are written into the memory 82 so as to be referred to in the processing by the motion compensation processing circuit 89. However, the I picture is written into the memory 82 as it is after being decoded by the inverse DCT / inverse quantization circuit 84. The P picture that has been motion-compensated by the motion compensation processing circuit 89 is written to the memory 82.

The motion compensation processing circuit 89 performs a motion compensation process based on the supplied signal with reference to the I picture and the P picture written in the memory 82 as described above. This motion compensation processing is decoding corresponding to the motion detection / compensation processing in the above-described compression encoding. The output of the motion compensation processing circuit 89 is composed of each frame completely decoded as an image signal. However, as described above, at the time of compression encoding, rearrangement is performed in order to perform motion compensation inter-frame prediction encoding. Therefore, the order of these frames is different from the order in the original image signal. Therefore, in order to restore the order of the frames in the decoded image signal, the frames are once written in the memory 82 and then extracted in the order in the original image signal.

The processing for restoring the order of the frames described above will be specifically described. Motion compensation processing circuit 8
The output of 9 is supplied to a signal compression circuit 91. The signal compression circuit 91 compresses the supplied signal. Then, the compressed signal is written into the memory 82 via the memory controller 86. The signals written in the memory 82 are taken out in the order of frames in the original image signal in accordance with an instruction from the controller 88. In such extraction, synchronization with a signal other than an image signal such as an audio signal is also considered. The extracted signal is transmitted to a signal expansion circuit 90.
And the original image signal is restored. The restored image signal is output to the subsequent stage via the post-processing unit 83.

After the decoding corresponding to the compression encoding is completed, it is necessary to return the order of the frames in the decoded image signal to the order of the frames in the original image signal. When writing to / removing from the memory 82, the signal compression circuit 91 / signal expansion circuit 90
Compression / expansion by the memory 8
2 can be reduced.

The fifth embodiment of the present invention described above is an example in which the present invention is applied to a decoding device that performs decoding corresponding to the compression encoding performed by the first embodiment. Similarly, the fifth embodiment of the present invention can be applied to a decoding device that performs decoding corresponding to the compression encoding performed by the second and third embodiments of the present invention. is there. Further, for example, a decoding device that performs decoding corresponding to a compression encoding device that performs compression encoding on a still image in accordance with JPEG or the like, such as the fourth embodiment of the present invention, It is also possible to apply the decoding device of the fifth embodiment. In this case, a configuration for supporting the motion compensation processing is not required.

The present invention is not limited to the above-described embodiment, and various applications and modifications can be considered without departing from the gist of the present invention.

[0081]

As described above, the present invention provides an MPE
In a compression encoding device and a decoding device used in a system for transmitting an image signal such as a moving image or a still image in accordance with regulations of G, JPEG, etc., a main compression encoding / decoding performed in accordance with each regulation. Is another compression /
It is designed to expand. For this reason, it is possible to reduce the capacity of the memory required for processing by the compression encoding device and the decoding device, and it is possible to reduce the cost of the entire system. Further, by adjusting the compression ratio and the like in the compression / decompression different from the main compression encoding / decoding, the capacity of the memory can be reduced so as to conform to the quality of the transmitted image and the like. .

Further, as described in the second embodiment of the present invention, control is performed so that compression / expansion different from main compression encoding / decoding is performed only on a specific portion of a signal. In such a case, the compression ratio or the like can be set high for the specific portion. In this case, the part that does not perform compression / decompression other than the main compression encoding / decoding is the part in the signal that is subjected to the feedback processing, that is, the main compression encoding performed on other parts in the signal. / A part referred to as a reference in decoding. By such control, the part used as a reference is compressed / decoded separately from the main compression encoding / decoding.
It does not include errors caused by stretching. For this reason, it is possible to prevent the quality of the transmitted image from deteriorating even if the compression ratio or the like of the specific portion in the signal described above is set high.

Further, as described in the third embodiment of the present invention, another compression / decompression other than the main compression encoding / decoding is performed by a unit equal to or less than an access unit in the main compression encoding / decoding. In the case where the compression / decompression is performed for each predetermined unit, compression / expansion different from the main compression encoding / decoding can be performed without deteriorating the access efficiency in the main compression encoding / decoding. it can. Therefore, MPE
While maintaining the original access efficiency of the system for transmitting image signals and the like in accordance with the regulations of G, JPEG, etc.
The capacity of the memory provided in association with the decoding device can be reduced.

On the other hand, as shown in the fifth embodiment of the present invention, even when the present invention is applied to the decoding device on the receiving side of the transmission system, the capacity of the memory provided in association with the decoding device Can be reduced. That is,
After the decoding corresponding to the compression encoding is completed, it is possible to reduce the memory capacity required to return the order of the frames in the decoded image signal to the order of the frames in the original image signal. it can.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a first embodiment of the present invention.

FIG. 2 is a schematic diagram for describing control performed according to a picture type in the first embodiment of the present invention.

FIG. 3 is a block diagram showing in detail a partial configuration of the first embodiment of the present invention.

FIG. 4 is a block diagram showing in detail another partial configuration of the first embodiment of the present invention.

FIG. 5 is achieved by a first embodiment of the present invention;
FIG. 4 is a schematic diagram for explaining a reduction in memory capacity.

FIG. 6 is a schematic diagram for explaining a case where the present invention is not applied, for comparison with FIG. 5;

FIG. 7 is a schematic diagram illustrating control performed according to a picture type in a second embodiment of the present invention.

FIG. 8 is a schematic diagram for explaining general DPCM compression.

FIG. 9 shows a DP performed in the first embodiment of the present invention.
FIG. 4 is a schematic diagram for explaining CM compression.

FIG. 10 is a diagram showing a D made in a third embodiment of the present invention.
FIG. 4 is a schematic diagram for explaining PCM compression.

FIG. 11 is a schematic diagram for describing control performed when the present invention is applied to a compression encoding / decoding device used in a system for transmitting a moving image.

FIG. 12 is a schematic diagram for explaining control performed when the present invention is applied to a compression encoding / decoding device used in a system for transmitting a still image.

FIG. 13 is a block diagram showing a configuration of a fourth embodiment of the present invention.

FIG. 14 is a block diagram showing a configuration of a fifth embodiment of the present invention.

FIG. 15 is a schematic diagram for explaining conventional compression encoding / decoding in an image transmission system such as MPEG or JPEG.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Encoder, 2 ... Memory, 3 ... Preprocessing part, 4 ... Discrete cosine transform (DCT) / quantization circuit, 5 ... Huffman coding circuit, 6 ... Memory controller , 7 ... code interface, 8 ... controller, 9 ... motion detection / compensation processing circuit, 10
..Inverse discrete cosine transform (inverse DCT) / inverse quantization circuit
11 ... signal compression circuit, 12 ... signal expansion circuit, 4
1 ... quantizer, 42 ... flip-flop, 43
... Inverse quantizer, 44 ... Switch, 45 ... Switch, 46 ... Subtractor, 47 ... Adder, 48 ...
..Multiplexer, 110 switching control circuit, 6
1 ... demultiplexer, 62 ... inverse quantizer, 6
3 ... adder, 64 ... switch, 65 ... flip-flop, 120 ... switching control circuit, 81 ...
· Decoder, 82 · · · memory, 83 · · · post-processing unit,
84: inverse discrete cosine transform (inverse DCT) / inverse quantization circuit, 85: Huffman decoding circuit, 86: memory controller, 87: code interface, 8
8 ... controller, 89 ... motion compensation processing circuit,
90 ... Signal expansion circuit, 91 ... Signal compression circuit

Claims (32)

[Claims] 1. A compression encoding apparatus for encoding an image signal, comprising: a memory; a compression circuit for compressing an input image signal; an expansion circuit for expanding a signal compressed by the compression circuit; And a memory control circuit for controlling writing and reading of the signal compressed by the compression circuit, and a coding circuit for coding the image signal decompressed by the decompression circuit. 2. The encoding circuit according to claim 1, wherein the encoding circuit detects a motion of the image signal, and calculates a difference image signal between an image signal to be encoded and a referenced image signal based on the detected motion information. A compression encoding apparatus comprising: a motion detection / motion compensation processing circuit that performs compression encoding; and a compression encoding circuit that compresses and encodes the difference image signal. 3. The encoding circuit according to claim 2, wherein the encoding circuit includes a local decoding circuit that locally decodes a signal compression-encoded by the compression encoding circuit, and the memory control circuit includes the local decoding circuit. The image signal locally decoded by the decoding circuit is stored in the memory, and the locally decoded image signal stored in the memory is supplied to the motion detection / motion compensation processing circuit in order to calculate a difference image signal. Compression encoding apparatus characterized in that: 4. The compression code according to claim 1, wherein said compression circuit and said decompression circuit execute a compression process and an expansion process on all image signals of the input image signal. Device. 5. The method according to claim 1, wherein the compression circuit and the expansion circuit perform a compression process and an expansion process on only a predetermined image signal of the input image signals. Compression encoding device. 6. The compression encoding apparatus according to claim 5, wherein the predetermined image signal is an image signal that is not locally decoded by the local decoding circuit. 7. The compression encoding circuit according to claim 6, wherein the compression encoding circuit encodes the image signal according to an encoding method defined by the MPEG standard, and the predetermined image signal is a B picture defined by the MPEG standard. A compression encoding device, comprising: 8. The compression circuit and the expansion circuit according to claim 7, wherein the compression circuit and the expansion circuit execute the compression processing and the expansion processing for the B picture with higher efficiency than the I picture and the P picture defined by the MPEG standard. Compression encoding apparatus characterized by the following. 9. The input image signal according to claim 1, wherein the input image signal includes a luminance signal and a chroma signal, and the compression circuit and the expansion circuit execute a compression process and an expansion process on the chroma signal. Compression encoding apparatus characterized by the following. 10. The compression circuit and expansion circuit according to claim 1, wherein the compression circuit and the expansion circuit perform a compression process and an expansion process on the image signal by calculating a difference between a pixel to be coded and a pixel adjacent to the pixel to be coded. Compression encoding apparatus characterized in that: 11. The encoding circuit according to claim 1, wherein the encoding circuit encodes the image signal in a predetermined block unit, and wherein the compression circuit and the decompression circuit determine a pixel to be encoded in a unit of the predetermined block unit or less. A compression coding apparatus characterized in that a compression process and a decompression process are performed on an image signal by calculating a difference between a pixel to be coded and a pixel adjacent thereto. 12. A decoding device for decoding an encoded image signal, comprising: a memory; a decoding circuit for decoding the encoded image signal; and a decoding circuit for storing in the memory. A compression circuit for compressing the compressed image signal; a decompression circuit for decompressing the signal compressed by the compression circuit; and a memory control for controlling writing and reading of the signal compressed by the compression circuit to and from the memory. And a decoding device. 13. The coded image signal according to claim 12, wherein the coded image signal includes motion vector information, and the decoding circuit decodes the coded image signal, and decodes the coded image signal. Or a decompression decoding circuit that obtains a difference signal, and a generation circuit that adds the difference image signal and a referenced image signal based on the motion vector information to generate a decoded image signal. A decoding device characterized by the above-mentioned. 14. The apparatus according to claim 13, wherein the compression circuit and the expansion circuit perform a compression process and an expansion process on all the image signals of the decoded image signal. Decoding device. 15. The image signal according to claim 12, wherein the decoded image signal includes a luminance signal and a chroma signal, and the compression circuit and the expansion circuit execute a compression process and an expansion process on the chroma signal. A decoding device characterized in that: 16. The compression circuit and the expansion circuit according to claim 12, wherein the compression circuit and the expansion circuit calculate a difference between a pixel to be coded and a pixel adjacent to the pixel to be coded, thereby performing a compression process and a decompression process on the image signal. A decoding device characterized in that the decoding is performed. 17. An encoding method for compressing and encoding an image signal, comprising: a compression step of compressing an input image signal by a predetermined compression method; a step of writing the compressed signal into a predetermined memory; Reading out the compressed signal stored in the memory of the above, expanding the read out signal, and encoding the transmitted image signal for transmission. Characteristic encoding method. 18. The image processing method according to claim 17, wherein the encoding step detects a motion of the image signal, and calculates a difference image signal between an image signal to be encoded and a referenced image signal based on the detected motion information. A coding method, comprising: a motion detection / motion compensation processing step of calculating; and a compression coding step of compression coding the difference image signal. 19. The compression encoding step according to claim 18, wherein the compression encoding step includes: a local decoding step of locally decoding a signal compression-encoded by the compression encoding step; The motion detection / motion compensation processing step reads the locally decoded image signal from the predetermined memory, and refers to the read locally decoded image signal. ,
An encoding method comprising the step of calculating the difference image signal. 20. The encoding method according to claim 19, wherein the compression step performs a compression process on all image signals of the input image signal. 21. The encoding method according to claim 19, wherein the compression step performs a compression process only on a predetermined image signal among the input image signals. 22. The encoding method according to claim 21, wherein the predetermined image signal is an image signal that is not locally decoded. 23. The compression encoding step according to claim 22, wherein the compression encoding step encodes an image signal according to an encoding method defined by the MPEG standard, and the predetermined image signal is a B picture defined by the MPEG standard. An encoding method, comprising: 24. The method according to claim 23, wherein the compression step is performed according to an I / O standard defined by the MPEG standard.
An encoding method characterized in that the B picture is compressed at a higher efficiency than a picture and a P picture. 25. The method according to claim 17, wherein the input image signal includes a luminance signal and a chroma signal, and the compression step performs compression on the chroma signal. Encoding method. 26. The method according to claim 17, wherein the compression step is configured to perform a compression process on the image signal by calculating a difference between a pixel to be encoded and a pixel adjacent to the pixel to be encoded. Characteristic encoding method. 27. The encoding method according to claim 17, wherein the encoding step encodes the image signal in a predetermined block unit, and the compression step encodes the image signal in a unit equal to or less than a block unit encoded in the encoding step. An encoding method characterized in that an image signal is compressed by calculating a difference between a pixel and a pixel adjacent to the pixel to be encoded. 28. A decoding method for decoding an encoded image signal, comprising: a decompression decoding step for decoding an encoded image signal; and a decoding method for storing the decoded image signal stored in a predetermined memory. A compression step of performing compression processing by a predetermined compression method; a step of writing a compressed signal to a predetermined memory; a step of reading a compressed signal stored in the predetermined memory; and a step of decompressing the read signal A decompression step for processing. 29. The coded image signal according to claim 28, wherein the coded image signal includes motion vector information, and the decoding step includes: decoding the image signal or the difference image from the coded image signal. Obtaining a signal, and, based on the motion vector information, adding the difference image signal and a referenced image signal to generate a decoded image signal. Method. 30. The decoding method according to claim 29, wherein in the compression step, a compression process is performed on all the image signals of the decoded image signal. 31. The method according to claim 28, wherein the decoded image signal includes a luminance signal and a chroma signal, and the compression step performs a compression process on the chroma signal. Characteristic decoding method. 32. The image processing method according to claim 28, wherein the compression step is configured to compress the image signal by calculating a difference between a pixel to be coded and a pixel adjacent to the pixel to be coded. Characteristic decoding method.