JP2001094991A - Data compressor, data compression method and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To compress data with a higher compression rate without, degrading quality of reproduced data in the case of compressing the data through vector quantization. SOLUTION: Prior to vector quantization processing, where a data stream having at least one or more data is divided into blocks to obtain vectors, a code vector similar to an extracted vector from a compression object is sought from a prepared code book in advance, and a code corresponding thereto is outputted, data of an odd numbered field or an even numbered field are eliminated from original image data which are a compression object consisting of data of odd numbered and even numbered fields adjacent to each other, and vector quantization is applied to only the other data so as to reduce the data quantity through the elimination of field information and defective retrieval of a code vector on the basis of flickering or the like, caused between the odd and even numbered fields is suppressed so as to enhance the image quality of the reproduced image.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data compression device, a compression method, and a storage medium storing a program for executing these processes, and more particularly, to a compression device using vector quantization as one of data compression techniques. It is suitable for use as a compression method and a storage medium.

[0002]

2. Description of the Related Art Conventionally, various data compression techniques have been proposed. Among them, a method called “vector quantization” is often used as one of data compression algorithms that can very easily perform decompression processing of compressed data. This algorithm has been known for a long time in the field of signal processing, especially for data compression of image signals and audio signals,
Or it has been applied to pattern recognition.

In this vector quantization, several pixel patterns of a certain size (for example, a block of 4 × 4 pixels) are prepared, and a unique number or the like is given to each of them (this set is called a “codebook”). ). Then, for example, blocks of the same size (for example, 4 × 4 pixels) are sequentially extracted from the image data of the two-dimensional array, a pattern most similar to the block is found in the codebook, and the pattern number is assigned to the block. Perform data compression. In vector quantization, a data sequence in one block corresponds to one code vector.

On the receiving side or decompressing side of the coded compressed data, the original image can be reproduced simply by extracting the pattern corresponding to the number for each block from the code book. Therefore, on the extension side,
If a codebook is received or stored in advance, no special calculation is required, and the original image can be reproduced with very simple hardware.

[0005]

In the above-described vector quantization, for example, when compressing data of an image or an audio signal, how to obtain a high quality reproduced image or reproduced audio while maintaining a high compression rate. Another problem is how to increase the compression ratio without lowering the quality of a reproduced image or the like.

Conventionally, in order to realize a high compression rate, for example, efforts have been made to increase the size of a block extracted from image data or to reduce the number of code vectors constituting a code book. However, for example, when compressing an image using a vector quantization technique, if a high compression ratio is to be realized, there arises a problem that the image quality of a reproduced image is necessarily deteriorated. Conversely, there is a problem in that an attempt to increase the image quality of the reproduced image does not result in an increase in the compression ratio.

SUMMARY OF THE INVENTION The present invention has been made to solve such a problem. When data compression is performed by vector quantization, the quality of reproduced data is not reduced.
A first object is to be able to compress data at a higher compression ratio. A second object of the present invention is to realize both a higher compression ratio and higher quality of data reproduced based on the compressed data obtained thereby.

[0008]

A data compression apparatus according to the present invention extracts image data of one of fields from original image data in which odd field data and even field data are alternately arranged. A field detection unit, and a data string having at least one or more data is divided into blocks to form a vector by compressing image data of one field extracted by the field detection unit, and the vector is prepared from a codebook prepared in advance. Vector quantization means for searching for a code vector similar to the vector extracted from the compression target and outputting a code corresponding thereto.

The data compression apparatus of the present invention has at least one
Vector quantization in which a data string having one or more data is divided into blocks to obtain a vector, a code vector similar to a vector extracted from a compression target is extracted from a prepared code book, and a corresponding code is output. Means for detecting a most frequent code from each code output from the vector quantization means, and replacing a code similar to the most frequent code among the output codes with a specific code. Means.

The data compression apparatus of the present invention has at least one
A vector sequence that blocks a data string having two or more data into a vector, searches for a code vector similar to a vector extracted from the compression target from a prepared codebook, and outputs a code corresponding to the vector. L, which performs LZSS dictionary compression on the code string output from the vector quantization means.
A ZSS compression unit is provided.

According to the data compression apparatus of the present invention, a flag byte header for discriminating predetermined specific data from other data based on a data string input as a compression target and indicating the positions of those data is provided. And a flag byte header converting means for sequentially performing a process of deleting the specific data from the data input as the compression target in units of a predetermined number of data strings.

The data compression apparatus of the present invention divides a data string having at least one or more data out of image data to be compressed into blocks, and extracts, for each block, a minimum value of each data in the block. A data dividing unit for dividing the data sequence in the block into a minimum value component and a component excluding the minimum value component by subtracting the extracted minimum value from each data in the block. The data string of the formation generated for each block by the respective vectors,
A vector quantization means for searching for a code vector similar to the vector of the formation from a code book prepared in advance, and outputting a code corresponding to the code vector, and separately determining the minimum value component and the formation To be processed.

According to the data compression method of the present invention, data in one field is deleted from original image data in which odd field data and even field data are alternately arranged, and then the data in the other field is compressed. As a target, a data string having at least one or more data is divided into blocks to obtain a vector, and a code vector similar to the vector extracted from the compression target is searched from a prepared code book, and a code corresponding to the vector is searched for. Is output.

According to the data compression method of the present invention, at least one
A data sequence having one or more data is divided into blocks to obtain a vector, a code vector similar to a vector extracted from a compression target is searched from a prepared code book, and a code corresponding to the vector is output. The most frequent code is detected from each of the output codes, and a code similar to the most frequent code among the output codes is replaced with a specific code.

According to the data compression method of the present invention, at least one
A data string having at least one data is divided into blocks to form vectors, a code vector similar to a vector extracted from the compression target is searched for from a prepared code book, and a code corresponding to the vector is output. LZSS dictionary compression is performed on the output code string.

According to the data compression method of the present invention, a data sequence having at least one data is divided into blocks and taken out from image data to be compressed, and the minimum value of each data in the block is extracted for each block. At the same time, by subtracting the extracted minimum value from each data in the block, the data sequence in the block is divided into a minimum value component and a formed component excluding the minimum value component. Each of the formed data strings is a vector, and a code vector similar to the formed vector is searched for from a previously prepared code book, and a code corresponding thereto is output.
The minimum value component and the formed component are separately processed.

The storage medium of the present invention is a computer-readable storage medium storing a program for causing a computer to function as each means of the above-described data compression device.

The storage medium of the present invention is a computer-readable storage medium storing a program for causing a computer to execute the procedure of the above data compression method.

[0019]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Some embodiments of the present invention will be described below in detail with reference to the drawings. In each of the embodiments described below, a color still image is taken as an example of data to be compressed, and the compression ratio of image data and the reproduction quality are further improved.

(First Embodiment) First, a first embodiment of the present invention will be described with reference to FIGS. First, the principle of image compression and decompression by vector quantization will be described with reference to FIG. As shown in FIG. 3, an original image 1 input as data to be compressed is configured by a large number of elements called pixels. Each pixel has information such as a luminance signal (Y signal) converted from an RGB signal and a color signal (U, V signal).

An input image block (macro block) 2 is obtained by extracting a block composed of a plurality of pixels from the input image 1. In the example of FIG.
Although 4 × 4 pixels are selected as the size of, this size may be anything. Since the input image block 2 has a plurality of pixels as described above, the luminance signal values and the chrominance signal values of each pixel can be collected and used as vector data. This is the input vector data.

Due to human visual characteristics, some input image blocks in the input image 1 may look almost the same in appearance. Such a plurality of input image blocks that look the same can be represented by a smaller number of image blocks. The image block codebook 3 has a plurality of image blocks (code vector data) representing a large number of input image blocks on the input image 1.
Code vector data is stored in the image block code book 3
The luminance and chromaticity of each pixel of the image blocks in the image data are vector data.

In the vector quantization, the entire input image 1 is divided into image blocks, and each image block 2 is used as input block data, and code vector data similar to the input vector data is searched from the code book 3. Then, the image can be compressed by transferring only the corresponding code vector data number. In order to reproduce the compressed image and obtain the reproduced image 4,
The code vector data corresponding to the transferred number may be read from the code book 3 and applied to the image.

Next, the operation regarding the Y, U, and V data performed at the time of the above-described vector quantization will be briefly described. In the following operation on Y, U, and V data according to the present embodiment, when performing data compression by vector quantization on a color image, the compression ratio on the compression side and the image quality of the reproduced image on the expansion side are determined. It is something that can be further improved.

FIG. 1 is a block diagram showing a configuration example of a data compression device according to the present embodiment. In FIG. 1, color original image data input from an image input unit 5 is temporarily stored in an image memory 6, read out by an address control or the like from a reading unit 7, and output. The color image data read from the image memory 6 is provided to a Y, U, V signal separation unit 8.

The Y, U, V signal separation section 8 separates a Y signal (luminance signal), a U signal, and a V signal (color signal) from the original image data. The blocking unit 9 includes the Y,
Each image data separated by the U and V signal separation unit 8 is
For example, it is divided into macroblocks of 4 × 4 pixels. At this time, the blocking unit 9 notifies the codebook storage unit 10 of which of the Y, U, and V signals is being blocked as a signal type designation signal.

The correlation calculation unit 11 calculates the similarity between each of the Y, U, and V signals separated and extracted in this manner with each of the code vectors in the codebook storage unit 10 for each macro. It is calculated for each block. Then, the code determination unit 12 outputs the code number corresponding to the code vector having the highest similarity as a compressed code (Winner code) of the macroblock based on the calculation result of the similarity by the correlation operation unit 11. . This allows
Vector quantization is performed.

At the time of executing such vector quantization, the correlation calculation unit 11 performs vector quantization using a codebook stored in the codebook storage unit 10 in advance. Which codebook to use is determined according to the signal type designation signal. That is, the codebook for the Y signal, the codebook for the U signal, and the codebook for the V signal are stored in the codebook storage unit 10 in advance, and the codebook for each signal is stored in a block. It is used sequentially according to the signal type output from the conversion unit 9 to the correlation degree calculation unit 11.

Generally, the most important signal that determines the image quality is
This is a Y signal (luminance signal). Therefore, in this embodiment,
A higher compression rate is assigned to the Y signal than to the UV signal. For example, the compression rates of the Y, U, and V signals are set to Y: U: V = 4: 1: 1. That is, for the Y signal, the vector quantization is performed using the signals of all the pixels as they are, whereas for the UV signal, the signal of the pixels is used as it is at a rate of one in four. That is, 4
The signals of three pixels of one pixel are thinned out and used. However, in the present embodiment, the compression rate of the UV signal is set to 4: 1: 1 in consideration of all pixels, instead of simply thinning out the UV signal.

As described above, when the compression rate is Y: U: V =
When the compression is performed with the ratio of 4: 1: 1, the compression ratio for one image can be increased by the amount of thinning out the UV signal, as compared with the case where the vector quantization is simply performed.
At this time, since a higher compression rate is assigned to the Y signal that is more important for the image quality, deterioration of the image quality can be suppressed. In addition, since human visual characteristics have a relatively insensitive property to color errors, there is no particular problem with respect to color reproducibility even when the UV signals are thinned out. Moreover, since the thinning is performed in consideration of all the pixels of the UV signal, the quality of the reproduced image is improved as compared with the case of the simple thinning.

That is, the data compression method according to the first embodiment is a method of compressing the amount of image data in the above-described Y, U, and V signals while suppressing deterioration in image quality.

FIG. 2A shows a data compression method in the case of line sampling in which a macroblock is taken in a line shape. Here, data compression by simple thinning is also shown as a comparative example. As shown in FIG.
In the data compression method according to the present embodiment, among the YUV signals, the data at the positions A to D is used as it is for the Y signal, and the average value of the four data at the positions A to D is calculated and used for the U and V signals. Thus, data compression is performed. That is, of the U and V signals, data corresponding to the positions of A, B, C, and D are averaged, and the average value is subjected to compression by vector quantization.

On the other hand, in the case of the simple thinning out of the comparative example, only the data at the position A is used for the vector quantization, and the data at the positions B, C and D are deleted. On the other hand, in the method of the present embodiment, the data at the positions B, C, and D are also considered as the average value. Therefore, although the data compression ratio is not different from simple thinning, data at all positions A, B, C, and D can be left, so that the image quality of a reproduced image can be further improved.

FIG. 2B shows a data compression method in the case of block sampling in which a macroblock is formed in a rectangular shape. In the case of block sampling, as in the case of line sampling, the data at the respective positions of A, B, C, and D are averaged and left, so that the reproduced image is maintained while maintaining the same compression ratio as in the case of simple thinning. Image quality can be improved.

Here, when obtaining the average value of the U and V data, the data at a specific position may be weighted. For example, in FIGS. 2A and 2B,
Instead of the average value (A + B + C + D) / 4, (2A + B +
The data at the position A may be weighted by calculating (C + D) / 5. If the data at location A is more important than the data at locations B, C, and D,
By performing such weighting, the image quality can be further improved. Further, in addition to such an average value and a weighted average value, a weighted addition value and other feature amounts may be calculated from the data at the positions A to D.

FIG. 2C is a block diagram showing the configuration and data flow of the data compression device according to the first embodiment. The configuration shown in FIG. 2C is provided between the blocking unit 9 and the correlation operation unit 11 in the example of FIG. The data blocked by the blocking unit 9 in FIG.
The input data is stored in the input data storage unit 13 in FIG. The decimation unit 14 outputs the Y signal as it is, but outputs 4 for the UV signal.
An average value or the like of the input data is calculated for each pixel, and the calculated value is output as a representative value of four pixels. The data output by the thinning operation unit 14 is sent to the generated data storage unit 15 and stored and held. Then, the image data stored in the generated data storage unit 15 is added to the correlation operation unit 11 shown in FIG.
Performs vector quantization.

As described above, according to the first embodiment of the present invention, the data amount of the U, V data among the Y, U, V data is reduced, and the average value and the weight are reduced at the time of the reduction. By using the average value or the like, it is possible to increase the compression ratio of the image data while minimizing the deterioration of the image quality of the reproduced image. In the present embodiment, the case where the configuration shown in FIG. 2C is used in combination with the vector quantization device is described as an example, but it may be applied to a compression method other than the vector quantization. Alternatively, the configuration shown in FIG. 2C may be used alone.

(Second Embodiment) Next, a second embodiment of the present invention will be described with reference to the drawings. In the second embodiment, the information amount of the original data is reduced by half by extracting only one of the odd and even lines of the odd and even lines of the input image before vector quantization. Then, vector quantization is performed only on the extracted field image.

FIG. 4 is a diagram for explaining the compression principle according to the second embodiment in comparison with a conventional example. Conventionally, as shown in FIG. 4A, a macroblock of 4 × 4 pixels has been extracted from an input image in which odd fields and even fields are alternately arranged. In this case, the data of the odd field and the data of the even field are included in the macro block. Therefore, vector quantization is performed for each macroblock including the data of the odd and even fields.

However, for example, when the even-numbered and odd-numbered fields of the NTSC signal are merged and subjected to vector quantization processing as an interlaced image, an image having a flicker (temporal shift) between the even-numbered and odd-numbered field lines can be obtained. Vector quantization will be performed. For this reason, a code vector including such a flicker element is extracted as having a high degree of similarity, which causes a defect search, and the flicker is emphasized by vector quantization. Therefore, the quality of an image reproduced from the compressed data deteriorates.

On the other hand, in the present embodiment, FIG.
As shown in (b), a field process of deleting one of the odd field and the even field is performed, and a macroblock of 4 × 4 pixels is extracted from only one of the odd field and the even field. By performing vector quantization on a macro block including only the data of the one field, not only the information amount can be halved, but also occurrence of a defective search due to flicker between odd and even fields can occur. And the deterioration of the image quality can be minimized.

FIG. 5 is a block diagram showing a configuration example of a data compression device according to the second embodiment. As shown in FIG. 5, image data input from the image input unit 20 is provided to an odd (even) field extraction unit 21. The odd (even) field extraction unit 21 extracts only one of the data of the odd field and the even field. Then, the data of either the extracted odd field or even field is stored in the odd (even) field data storage unit 22.

Thereafter, vector quantization is performed on the data stored in the odd (even) field data storage unit 22 by referring to the codebook in the codebook storage unit 24 in the vector quantization unit 23. The code number sequence for each macro block generated by the vector quantization is stored in the code number sequence storage unit 25. Although only the field processing is described in the present embodiment, when the averaging and weighting processing of the U and V data described in the first embodiment is performed, one of the extracted data after the field processing in the present embodiment is used. After the RGB data of the field No. is converted into the YUV data, it is performed before the vector quantization.

When data compressed by vector quantization involving the above-described field processing is reproduced, for example, when processing is performed to delete even fields and leave odd fields, a code compressed for odd fields is used. The original image can be reproduced by extracting the corresponding code vector from the number sequence from the code book, arranging it at the position of the original image, and obtaining an image corresponding to the even field between the odd fields by interpolating. When the compression target is a still image, if there is image data of one field, the original image can be sufficiently reproduced. Rather, the image quality is improved by suppressing a defect search at the time of vector quantization.

As described above, according to the second embodiment of the present invention, one of the odd field and the even field is deleted before performing the vector quantization, thereby reducing the image data amount. By performing vector quantization only on the data of one of the adopted fields, it is possible to suppress a defective search based on flicker occurring between an odd field and an even field, and to reduce the image quality of the reproduced image. Can be improved.

(Third Embodiment) Next, a third embodiment of the present invention will be described. The third embodiment is a method of further compressing data after vector quantization, and performs LZSS dictionary compression on the code number sequence after vector quantization and compression to apply further data compression. Since the LZSS dictionary compression is completely lossless compression, it is possible to return to a code number sequence by vector quantization before LZSS dictionary compression during reproduction.

Here, the LZSS dictionary compression refers to a process in which past data is referred to as a dictionary and encoded using the longest matching data string in the dictionary. , The compression effect increases. The LZSS dictionary compression will be briefly described with reference to FIG. FIG. 6A shows a data string before LZSS dictionary compression. In the data string before LZSS dictionary compression, the second to sixth data strings (BCDEF) and the seventh to eleventh data strings (BCDEF) are the same data.

The LZSS dictionary compression is a method of expressing the same data string as the data string appearing earlier using the preceding data string as a dictionary. In this example, the seventh to eleventh data strings ( BCDEF) is described using the second to sixth data strings (BCDEF) as a dictionary.

FIG. 6B shows a data string after LZSS dictionary compression. The seventh to eleventh data strings (BCDEF) are the same as the second to five consecutive data strings, and are therefore described as (2,5). Normally, one data (for example, the first A) is composed of one byte, but the data amount of (2, 5) in FIG. 6B is two bytes. Therefore, the seventh to eleventh data strings (BC
DEF) originally required 5 bytes of data, but now only needs 2 bytes, and can perform data compression of 3 bytes.

FIG. 6C shows a 2-byte data string storing (2, 5). Here, "2" indicating the head position of a data string referred to as a dictionary is represented by 12-bit data, and "5" indicating the number of data contained in the dictionary is represented by 4-bit data. You. When the data representing the number of the longest matching data in the dictionary is 4 bits long, a data string of 0 to 15 can be stored as a dictionary as the number of continuous data. By adding a predetermined value to the bit-length data value and expressing it, for example, it is possible to store more than 1 to 16 data strings such as 4 to 19 data strings as a dictionary.

On the other hand, in the vector quantization, each macro block extracted from the data to be compressed is represented by any one of the limited code vectors in the codebook. In many cases, the same code number sequence is output for the portion where the code has been set. Therefore, since the code number sequence after vector quantization is expected to output many data strings of the same phrase, when the LZSS dictionary compression is performed on this code number sequence, it is stored as a dictionary. Many data strings can be obtained. Therefore, the data string after vector quantization and the LZSS dictionary compression have high affinity, and data compression with a high compression ratio can be performed.

Next, the configuration and data flow of the data compression apparatus according to this embodiment will be described with reference to FIG. As shown in FIG. 7, original image data to be compressed is input to the input data storage unit 30. Then, the vector quantization section 31 performs vector quantization on the input data. Here, as described in the first embodiment, the similarity between each code vector constituting the codebook stored in the codebook storage unit 32 and each macroblock extracted from the input data is determined for each macroblock. , And sequentially obtaining code numbers corresponding to the most similar code vectors, a code number sequence is created.

The created code number sequence is stored in the code number sequence storage unit 33. Then, the code number sequence stored in the code number sequence storage unit 33 is
The compression is further performed by performing the LZSS dictionary compression processing in the ZSS compression unit 34, and the compressed image data is sent to the generated data storage unit 35.

As described above, according to the third embodiment of the present invention, the data sequence after vector quantization is
Since LZSS dictionary compression is performed using the fact that there are many data strings of the same phrase, a large number of data strings to be obtained as a dictionary can be obtained, and the data compression rate can be increased.

(Fourth Embodiment) Next, a fourth embodiment of the present invention will be described. The fourth embodiment is a method of compressing a code number sequence after vector quantization as in the third embodiment.

In the fourth embodiment, the data sequence after vector quantization is replaced with the most frequent code vector for a data sequence similar to the most frequent code vector. As a result, more data strings of the same phrase are generated, and the compression effect when LZSS dictionary compression or the like is performed thereafter can be further enhanced.

FIG. 8 is a diagram showing a configuration example of a data compression device according to the fourth embodiment. As shown in FIG. 8, the original image data stored in the input data storage unit 40 is sent to the vector quantization unit 41, and the vector quantization is performed using the codebook in the codebook storage unit 42. The code number sequence generated here is stored in the code number sequence storage 43.
Is stored.

Thereafter, the code number string in the code number string storage section 43 is sent to the most frequent code number detection section 44, and the most frequently used code number is selected from among the code numbers generated from the original image data. Is specified. For example, when a still image including a background is used as original image data to be compressed, a code number corresponding to the background image is detected as the most frequently occurring code. The detected most frequent code number is sent to the similarity determination unit 45. Similarity determination unit 45
Calculates the similarity between the code vector corresponding to each code number stored in the code number string storage unit 43 and the code vector corresponding to the most frequent code number detected by the most frequent code number detection unit 44. I do. And
A code number having a similarity greater than a predetermined value is replaced with a most frequently occurring code number. The code number sequence generated by the similarity determination unit 45 replacing the corresponding code number with the most frequently occurring code number is sent to the generated data storage unit 46. The LZSS compression unit 47 is configured to store the generated data
LZSS dictionary compression is performed on the code number sequence stored in No. 6.

Since the similarity determination section 45 replaces only the code number similar to the code vector of the most frequently occurring code number, it is possible to minimize the deterioration of the apparent image quality. The data to be replaced is not limited to the most frequently occurring code numbers. For example, when the code number is sent in units of 2 bytes, it is possible to replace it with “00h”, “FFh”, etc., in addition to the above-mentioned most frequent code number. Note that when replaced with “00h”, the code number 0 has no meaning, but “FF”
h ”, the maximum code number value is 2
No code number is wasted if it is shorter than the byte length.

Hereinafter, the replacement code will be described with reference to FIG. 9A shows a case where the code number is replaced with the most frequently used code number, FIG. 9B shows a case where the code number is replaced with “00h”, and FIG. 9C shows a case where the code number is replaced with “FFh”.

In the present embodiment, when performing LZSS dictionary compression on a code number sequence in which code numbers have been replaced, in order to further enhance the compression effect, a data sequence obtained by collecting only the upper bytes of each code number is used. , And a data string that collects only the lower bytes. In this case, FIG.
As shown in (a), when replacing with the most frequently occurring code number, if the upper bits of the replaced code number include a discontinuous value, the length of the data string usable as a dictionary is reduced. The probability of shortening increases. On the other hand, FIG.
(B), as shown in FIG. 9 (c), when replacing with "00h" or "FFh", there is no discontinuous part in the replaced code.

Therefore, as shown in FIGS. 9 (b) and 9 (c), when the word is replaced with “00h” or “FFh”, the length of a continuous phrase usable as a dictionary is increased. Therefore, when performing LZSS dictionary compression thereafter, more data strings can be obtained as a dictionary and converted. Therefore, "00h" or "FF
When the replacement with h ″ and the LZSS dictionary compression are used together, data compression with a higher compression ratio can be performed.

As described above, according to the fourth embodiment of the present invention, among the code numbers after vector quantization, those having similar code vectors to the most frequently occurring code numbers are described. By performing the replacement with a predetermined code number, the bias of the same code number can be increased, and the compression effect when the LZSS dictionary compression processing is performed thereafter can be enhanced.

(Fifth Embodiment) Next, a fifth embodiment of the present invention will be described. In the fifth embodiment, the fourth
1 byte for the data string after processing according to the embodiment of FIG.
The code number sequence is sequentially compressed in units of eight using the header e (hereinafter, referred to as a flag byte header).

FIG. 10A shows eight code number strings after replacement with the most frequently occurring code numbers according to the fourth embodiment. Where code3, code
The code numbers of code 4, code 6, code 7 are replaced according to the fourth embodiment.
It is assumed that the code numbers 1, code2, code5, and code8 have not been replaced.
That is, code1, code2, code5, co
Each code number of de8 is a code number that could not be replaced because the corresponding code vector is dissimilar to the code vector corresponding to the most frequently occurring code number.

FIG. 10B shows a data string obtained by performing the flag byte header conversion processing of the present embodiment on the eight code number strings shown in FIG. 10A. In the data string after the flag byte header conversion processing, the code number replaced by the fourth embodiment is deleted,
A 1-byte flag byte header is added at the beginning.
On the other hand, the code numbers that have not been replaced are left behind the flag byte header.

Each bit of the flag byte corresponds to eight code numbers. By setting the bit value of the flag byte header to “0”, it is indicated that the corresponding code number has been deleted. That is, it is indicated that the code number corresponding to the bit position whose value is set to “0” in the flag byte header is the code number replaced according to the fourth embodiment. On the other hand, by setting the bit value of the flag byte header to “1”, it is indicated that a code number that has not been replaced exists in a location corresponding to the bit position.

For example, consider a case where a code number similar to the most frequently occurring code vector is replaced with “FFh”.
In this case, the replaced code number sequence is read in units of eight, and the upper byte of each code number is confirmed.
When the upper byte of each code number is other than “FFh”, the corresponding bit of the flag byte header is set to “1”,
In the case of "FFh", it is set to "0" (or vice versa), and a flag byte header is generated. Furthermore, if the upper byte is "F
By deleting the code number of “Fh”, FIG.
A data string as shown in FIG.

The flag byte header is a 1-byte data sequence, and one code number sequence after vector quantization compression is a 2-byte data sequence. If there is at least one replaced code number sequence, data can be compressed by performing flag byte header conversion. The data can be further compressed by using the data after compression by the flag byte header conversion and then performing LZSS dictionary compression or the like.

Next, a description will be given of a method of optimizing data remaining without being deleted due to the addition of the flag byte header by using the data after the flag byte header conversion.
FIG. 11 schematically shows a method of optimizing data after conversion of the flag byte header. For example, when one code number is represented by a 12-bit length, since data transmission and the like are performed in byte units, the upper 4 bits indicated by shading in the upper part of FIG. ing. In the present embodiment, the useless four bits are optimized using two consecutive code numbers remaining after the flag byte header conversion.

That is, as shown in FIG.
Since the first 4 bits of modeA are an unused data string that is not used, this unnecessary 4-bit data string and C
The data sequence of the upper 4 bits of modeA is exchanged. The first 4 bits of the next CodeB data string are also useless data strings, but as shown in FIG.
For deB, the data strings are not exchanged. Then, in this state, the data strings of the upper 8 bits of CodeA and CodeB are added.

As a result, as shown in FIG.
The data sequence is arranged in the order of the upper 4 bits of CodeA, the upper 4 bits of CodeB, the lower 8 bits of CodeA, and the lower 8 bits of CodeB, and optimization from a total of 4 bytes of data to 3 bytes of data is performed. In this way, by performing optimization by adding the data strings of CodeA and CodeB, it is possible to reduce useless data of a total of 8 bits. Therefore, FIG.
Code1, Code2, Code5, Code shown in
By performing this optimization processing on 8, the data amount of 9 bytes can be reduced to 7 bytes.

If the number of code numbers subordinate to one flag byte header is an odd number, the upper 4 bits of the last code number remain as blank useless data.

FIG. 12 is a schematic diagram showing a data string after optimization. At the head of the processed data, a flag byte header is arranged collectively, and after that, data is arranged in the order of a data group of lower bits and a data group of upper bits. When pairs of a flag byte header and a code number sequence subordinate to the flag byte header are sequentially arranged, the upper 4 bits are wasted every time there is a set in which the number of code numbers subordinate to one flag byte header is odd. Data
When viewed as a whole, useless data is scattered.

On the other hand, as shown in FIG. 12, the data after the flag byte conversion is classified into a flag byte header group, a lower bit group, and an upper bit group, and arranged by being packed from the end, thereby providing the flag byte header. Even if there are multiple pairs where the number of code numbers remaining after conversion is an odd number,
When the optimization process is performed, the last part (the part indicated by hatching in FIG. 12) of the upper-order bit group is wasted at most. Therefore, by adopting the arrangement shown in FIG. 12, the data compression ratio can be further increased. Also,
As described in the fourth embodiment, by arranging the data sequence of the lower bits and the data sequence of the upper bits respectively, the length of a continuous phrase usable as a dictionary can be increased. The compression effect when LZSS dictionary compression is performed later can be enhanced.

FIG. 13 is a block diagram showing the configuration and data flow of a data compression device according to the fifth embodiment.
In the data string storage unit 50, the data string after the replacement with the most frequently occurring code numbers according to the fourth embodiment (FIG. 1)
0 (see (a)). Data reading unit 5
1 reads eight code number strings from the data string storage unit 50 at a time, and sends them to the flag byte header generation unit 52 and the data conversion unit 53. The flag byte header generation unit 52 generates a 1-byte flag byte header according to whether each of the eight input code numbers has been replaced with a predetermined code. Also, the data conversion unit 5
In step 3, the replaced code number is deleted from the input eight code numbers. Then, optimization processing is performed on the code number sequence remaining at this time. The flag byte header generated by the flag byte header generator 52 and the lower bit group and the upper bit group of the code number generated by the data converter 53 are stored in a generated data storage section as shown in FIG. 54.

As described above, according to the fifth embodiment of the present invention, by performing the flag byte header conversion and compression, the code number sequence of the same value replaced by the fourth embodiment can be replaced by one. It becomes possible to perform data compression by converting to a byte flag byte header. Further, by optimizing the data string after the flag byte header conversion, further data compression can be performed.

In this embodiment, the case where the flag byte header conversion compression is used in combination with the vector quantization and the LZSS dictionary compression conversion has been described, but it is also possible to use it alone. In this case, if one data is a data string composed of two or more bytes, the data amount can be compressed by converting specific data in the data read in units of eight into a flag byte header of one byte. it can.

(Sixth Embodiment) Next, a sixth embodiment of the present invention will be described. The sixth embodiment is a combination of the first to fifth embodiments.

FIG. 14 is a block diagram showing a configuration example of a data compression device according to the sixth embodiment. As shown in FIG. 14, the RGB data of the original image is input from the image input unit 60 and sent to the odd (even) field detection unit 61. The odd (even) field detector 61 performs the field processing described in the second embodiment, and deletes data of either the odd field or the even field. Here, the data of the even field is deleted, and only the data of the odd field is extracted and sent to the odd field data storage unit 62.

The data stored in the odd field data storage unit 62 is then sent to the thinning operation unit 63. In the decimation unit 63, as described in the first embodiment,
Among the YUV data converted from the RGB data, the color signal (UV data) is subjected to a thinning operation using an average value, a weighted average value, or the like. The data on which the thinning operation has been performed is provided to the vector quantization unit 64, and the vector quantization is performed using the codebook in the codebook storage unit 75. The code number sequence output for each macroblock by the vector quantization is stored in a code number sequence storage unit 65.
Is stored.

The code number sequence stored in the code number sequence storage unit 65 is replaced by the most frequently occurring code number detection unit 66 and the similarity determination unit 67 to replace the data described in the fourth embodiment. Will be That is, the most frequent code number detection unit 66 detects the most frequent code number from the code numbers subjected to vector quantization, and
7 replaces the code number similar to the most frequent code number with the most frequent code or “00h” or “FFh”. The code number sequence thus replaced is sent to and stored in the code number sequence storage unit 68.

The code number string stored in the code number string storage section 68 is read by the data reading section 69 in units of eight, and given to the flag byte header generation section 70 and the data conversion section 71. Then, the flag byte header generation unit 70 and the data conversion unit 71 perform the flag byte header conversion processing described in the fifth embodiment. That is, the flag byte header generation unit 70
Based on the code number sequence in unit, a flag byte header is generated which distinguishes between the replaced code number and the non-replaced code number in the similarity determination unit 67 and indicates their respective positions.

On the other hand, the data conversion unit 71 performs conversion for deleting the code number replaced by the similarity determination unit 67, and performs optimization processing on the code number string remaining without being deleted. Done. Then, the generated flag byte header and the optimized code number sequence are stored in the generated data storage unit 72 in the form shown in FIG.

Finally, the data string stored in the generated data storage section 72 is subjected to a third
The LZSS dictionary compression described in the embodiment is performed. The data compressed by the LZSS dictionary compression is stored in the data storage unit 74, and then transmitted on a transmission path as necessary, or recorded on a data storage medium.

As described above, in the sixth embodiment, the first to first embodiments are described.
By combining all the data compression methods according to the fifth embodiment, it is possible to perform the data compression while maximizing the compression ratio while suppressing the image quality deterioration of the reproduced image. Of course, in order to simplify the device configuration, the data compression by any of the methods of the first to fifth embodiments may be omitted.

(Seventh Embodiment) Next, a seventh embodiment of the present invention will be described. FIG. 15 is a flowchart illustrating a procedure of a data compression method according to the seventh embodiment. Hereinafter, the data compression method according to the seventh embodiment will be described with reference to FIG.

In steps S70 to S72 shown in FIG. 15, the data compression described in the first and second embodiments is performed. That is, when the original image data represented by R, G, and B is input in step S70, the field processing described in the second embodiment is performed in step S71. As a result, data of one of the even field and the odd field is deleted, and only the data of the other field is extracted as a target of the subsequent compression processing.

Then, in step S72, a process of converting the R, G, B data of the field extracted in step S71 into Y, U, V data is performed. At this time,
The U and V data are thinned out based on the average value, the weighted average value, and the like described in the first embodiment.
The data is compressed so that the color signal value of the pixel is represented by one feature amount obtained from them.

In the next step S73, a 4 × 4 macroblock (MB) conversion process is performed. The 4 × 4 macroblock (MB) conversion process is an address conversion process for extracting data for each rectangular macroblock of 4 × 4 pixels from a data sequence arranged as each line of an image.

In the next step S74, the data in each macroblock extracted from the image is
A process of dividing into a formed component and a minimum value component is performed. Here, the minimum value component means a component having the minimum value among the pixel values included in one macroblock. Also, the formed component is
This value is obtained by subtracting the value of the minimum value component from each pixel value included in one macroblock.

FIG. 16 is a diagram schematically showing the formation of 4 × 4 macroblock data and the minimum value component. FIG.
Indicates the value of, for example, Y data among the Y, U, and V data. Here, the original 4 × 4 macroblock data 8
Among the 16 pixel values included in 5, the minimum value is “1”.
0 ". This is the minimum value component of this macroblock. Further, by subtracting the minimum value component from all the pixel values in the macroblock data 85, each pixel value in the macroblock is minimized. A formation part 86 represented only by an increment from the pixel value is generated, and the U and V data are similarly divided into a formation part and a minimum value component.

Then, the subsequent steps S75 to S79 and steps S82 to S86 separately process the formed components and the minimum value components thus divided. That is, the formed 4 × 4 macroblock 86
The processing procedure proceeds to step S76 via step S75 shown in FIG. 15, where the vector quantization compression of the formation is performed on each of the YUV data with reference to the codebook.

Then, in step S77, the most frequently occurring code numbers and their frequencies (Y: n2_y, U: n2_u, V:
n2_v) is detected for each of the YUV data. Here, since the target is the formation of a simple pattern generated by subtracting the minimum value component, similarity calculation such as Manhattan distance can be substantially unnecessary. Therefore, here, the number sequence obtained by the vector quantization is output as it is, without replacing the code number based on the similarity with the most frequently occurring code vector described in the fourth embodiment. Next, in step S78, it is determined whether or not to perform the flag byte header conversion described in the fifth embodiment. Here, based on the frequency of use of the most frequent code number detected in step S77, it is determined whether the flag byte header conversion is effective (whether the data amount is reduced as a whole). That is, the flag byte header conversion is intended to compress the data by adding a flag byte header instead of deleting the most frequently occurring code number. It is conceivable that the amount of data added is larger than the amount. Therefore, when the compression effect cannot be obtained even if the flag byte header conversion is performed, the flag byte header conversion is not performed. If it is determined that the flag byte header conversion is to be performed, a flag byte header conversion process is performed in step S79. If it is determined that the flag byte header conversion is not to be performed, the process proceeds to step S80 without performing the process of step S79.

On the other hand, the processing procedure for the minimum value component divided in step S74 proceeds to step S83 via step S82, where gradation quantization is performed on each of the YUV data. In the gradation quantization process,
For example, in the case where 256 gradations are represented by 8-bit data, the gradation is coarsened and reduced to 128 gradations.
Bit data can be reduced. In addition, 6 more
By lowering to 4 gradations, it is also possible to reduce to 6-bit data. Hereinafter, the value of the minimum value component that has been subjected to gradation quantization is referred to as a “gradation value”.

Although the gradation quantization is performed on the minimum value component in this case, considering that the minimum value component (scalar) is a one-dimensional vector, a vector quantization code Vector quantization may be performed with reference to a one-dimensional vector codebook having a smaller data amount than a book. Alternatively, a plurality of minimum value components may be combined to newly generate vector data, and vector quantization may be performed on the vector data.
For the minimum value component, data compression may be performed by a method other than gradation quantization, or processing may be simplified by not performing data compression. Even if data compression is not performed for the minimum component,
Since vector quantization is performed on the formed part, it is possible to achieve compression of the entire data.

Next, in step S84, the most frequently occurring gradation values and their frequencies (Y: n1_y, U: n1_u, V: n1)
_V) for each of the YUV data.
As in the case of the formation, the distance calculation such as the Manhattan distance is not necessary here, and the data string obtained by the gradation quantization is output as it is. Then, in step S85,
It is determined whether to perform the flag byte header conversion described in the fifth embodiment. In this step S85, whether or not to perform the flag byte header conversion is determined by Y, U,
The determination is performed based on whether or not the following expression holds for each of the V data.

[0098]

(Equation 1)

In each of the above equations, the left side shows the original data amount, and the right side shows the data amount after the flag byte header conversion. In addition, m indicates the number of macroblocks included in the image, and n1_y, n1_u, and n1_v indicate the number of appearances of the most frequently occurring gradation values of the YUV data.
When the above expression is satisfied in any of the Y, U, and V data, the data amount after the flag byte header conversion is smaller than the original data amount. For data that satisfies
At 85, it is determined that the flag byte header conversion processing is to be performed. If it is determined that the flag byte header conversion is to be performed, a flag byte header conversion process is performed in step S86. If it is determined that the flag byte header conversion is not to be performed, the process proceeds to step S80 without performing the process of step S86.

In step S80, the LZSS dictionary compression described in the third embodiment is performed on both the formed component and the minimum value component. After that, the data compressed in step S81 is output.

Next, a method of expanding data compressed by the data compression processing of FIG. 16 will be described with reference to FIG. FIG. 17 is a flowchart showing the procedure of data decompression.

First, in step S90, compressed data is input. In the following step S91, LZSS
Dictionary expansion is performed. This is an inverse transform of the LZSS dictionary compression described in the third embodiment. Next step S9
In step 2, reverse flag byte header conversion is performed. this is,
This is a reverse conversion of the flag byte header conversion described in the fifth embodiment. That is, the process of inserting the most frequently occurring data value deleted at the time of compression into the data position indicated by “0” in the flag byte header to sequentially create the original eight data strings, and deleting the flag byte header I do. As a result, the code number sequence output by vector quantization at the time of compression is reproduced in step S93 for the formed part, and the step S93 is performed for the minimum value component.
At 94, the gradation value sequence output by gradation quantization at the time of compression is reproduced.

In the processing in steps S93 to S96, the data is divided into the formed components and the minimum value components, and the respective components are subjected to decompression processing. Regarding the formed components, steps S93 to S
Proceeding to 94, a vector quantization decompression process is performed. For the minimum value component, steps S95 to S96
Then, the gradation quantization expansion processing is performed.

Next, in step S97, the formed data subjected to the vector quantization / expansion processing and the minimum value component data subjected to the gradation quantization / expansion processing are combined for each macroblock. You. That is, the original macroblock data before division is reproduced by adding the values of the expanded minimum value components to all the pixel values in the macroblock expanded for the formed part.

In the next step S98, by performing inverse macroblock conversion, the macroblock data of the 4 × 4 configuration is converted into a line. The processing so far is Y
Performed separately for each of the UV data. Then, in the next step S99, the Y, U, V data are R, G,
It is converted to B data.

In the next step S100, field interpolation processing is performed, and processing for interpolating data deleted during compression between even fields or between odd fields is performed. In the last step S101, the R, G, and B images reproduced by the above series of processing are output as reproduced images.

Although the decompression process has not been described in each of the first to sixth embodiments, in these embodiments as well, in order to reproduce the original data from the compressed data, it is necessary to use the data. Following the compression process in reverse,
By performing data decompression, a reproduced image can be obtained.

As described above, according to the seventh embodiment of the present invention, the data after Y, U, and V conversion is divided into a formed component and a minimum value component, and vector quantization is performed on the formed component. By doing so, it is possible to simplify the codebook used exclusively for the formation, compared to the codebook for the original image itself in which various features are intertwined, and it is necessary to perform vector quantization of the formation. The number of code vector patterns can also be reduced. As a result of reducing the number of code vectors, the compression ratio of data compression can be increased, and the storage capacity of the code book can be reduced.

Further, by performing vector quantization on the formed component from which the minimum value component has been subtracted, it is possible to reduce the luminance value having a large influence on image quality and perform vector quantization. Therefore, it is possible to prevent the change in the luminance value existing at the boundary of the macroblock from being emphasized by the vector quantization, and to obtain a reproduced image in which the data values at the boundary of the macroblock are smoothly connected. It becomes possible. Also, in this case, since the minimum value component obtained by dividing the formed component is separately compressed and the information is left, the decompression process is performed by using the compressed component.
Deterioration of the image quality of the reproduced image can be suppressed.

(Other Embodiments) Each functional block and processing procedure shown in the above various embodiments may be constituted by hardware, or may be constituted by CPU, MPU, R
It may be configured by a microcomputer system including an OM and a RAM, and its operation may be realized according to a work program stored in a ROM or a RAM. The present invention also includes a software program for realizing the functions described above, which is provided to a RAM so as to realize the functions of the respective functional blocks, and which is executed by operating the functional blocks according to the programs. include.

In this case, the software program itself realizes the functions of the above-described embodiments, and the program itself and means for supplying the program to a computer, for example, a recording medium storing the program Constitute the present invention. As a storage medium for storing such a program, the above-described ROM or RAM
Besides, for example, floppy disk, hard disk,
Optical disk, magneto-optical disk, CD-ROM, CD-
I, CD-R, CD-RW, DVD, zip, magnetic tape, nonvolatile memory card, or the like can be used.

When the computer executes the supplied program, not only the functions of the above-described embodiment are realized, but also the OS (operating system) or other application software running on the computer. Needless to say, such a program is also included in the embodiment of the present invention when the functions of the above-described embodiment are realized in cooperation with the above.

Further, after the supplied program is stored in the memory provided in the function expansion board of the computer or the function expansion unit connected to the computer, the program is provided in the function expansion board or the function expansion unit based on the instruction of the program. It is needless to say that the present invention also includes a case where the CPU or the like performs part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

[0114]

According to the first aspect of the present invention,
Since only one of the odd field and the even field is used, a higher compression rate can be obtained when data compression by vector quantization is performed, and the deterioration of image quality occurring between both fields is reduced. Deter,
It is possible to improve the quality of data reproduced based on the obtained compressed data.

According to the invention of claim 3 of the present invention, when performing data compression by vector quantization, by replacing the code with the most frequently occurring code, the similarity of each code after vector quantization is obtained. The degree can be increased. Therefore, data compression can be performed at a high compression rate thereafter.

According to the ninth aspect of the present invention, when performing data compression by vector quantization, LZSS dictionary compression is performed on each code after vector quantization, thereby improving the quality of reproduced data. It is possible to compress data at a higher compression ratio without dropping.

According to the tenth aspect of the present invention,
By calculating the pixel value of the color signal for each of a plurality of pixels, and representing the color signals of the plurality of pixels with the obtained calculated value,
Of the image data, it is particularly possible to compress color signal data.

According to the eleventh aspect of the present invention,
By generating a flag byte header indicating specific data and deleting the specific data, it is possible to increase the data compression ratio when a large number of specific data is included.

According to the twelfth aspect of the present invention,
By dividing the data string into a formed part and a minimum value component and processing both separately, it is possible to perform data compression suitable for each of the formed part and the minimum value component.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration example of a data compression device according to a first embodiment of the present invention.

FIG. 2 shows Y, U, and Y performed in the first embodiment of the present invention.
FIG. 5 is a schematic diagram for explaining an operation regarding V data.

FIG. 3 is a schematic diagram for explaining image compression by vector quantization in each embodiment of the present invention.

FIG. 4 is a schematic diagram for explaining data compression in a second embodiment of the present invention.

FIG. 5 is a block diagram illustrating a configuration example of a data compression device according to a second embodiment of the present invention.

FIG. 6 is a schematic diagram for explaining LZSS dictionary compression according to a third embodiment of the present invention.

FIG. 7 is a block diagram illustrating a configuration example of a data compression device according to a third embodiment of the present invention.

FIG. 8 is a block diagram illustrating a configuration example of a data compression device according to a fourth embodiment of the present invention.

FIG. 9 is a schematic diagram showing an example of a data string after replacement with the most frequently occurring code numbers according to the fourth embodiment of the present invention.

FIG. 10 is a schematic diagram for explaining data compression by flag byte header conversion according to a fifth embodiment of the present invention.

FIG. 11 is a schematic diagram for explaining a method of performing optimization using data after conversion of a flag byte header according to a fifth embodiment of the present invention.

FIG. 12 is a schematic diagram showing a data string after optimization according to a fifth embodiment of the present invention.

FIG. 13 is a block diagram illustrating a configuration example of a data compression device according to a fifth embodiment of the present invention.

FIG. 14 is a block diagram illustrating a configuration example of a data compression device according to a sixth embodiment of the present invention.

FIG. 15 is a flowchart showing a procedure of data compression according to the seventh embodiment of the present invention.

FIG. 16 is a schematic diagram illustrating an example of data of a formed component and a minimum component divided according to the seventh embodiment of the present invention.

FIG. 17 is a flowchart showing a procedure of data decompression according to a seventh embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Original image 2 Input image block 3 Codebook 4 Reproduction image 5, 20 Image input part 6 Image memory 7 Readout part 8 Y, U, V signal separation part 9 Blocking part 10, 24, 32, 42 Codebook storage part 11 Correlation calculation unit 12 Code determination unit 13, 30, 40 Input data storage unit 14 Decimation calculation unit 15, 35, 46, 54, 72 Generated data storage unit 21 Odd (even) field detection unit 22 Odd (even) field data storage Units 23, 31, 41, 64 Vector quantization units 25, 33, 43, 65, 68 Code number string storage units 34, 47, 73 LZSS compression units 44, 66 Most frequently occurring code number detection units 45, 67 Similarity determination units 50 Data string storage unit 51, 69 Data reading unit 52, 70 Flag byte header generation unit 53, 71 Data conversion unit 6 0 Image input unit 61 Odd (even) field detection unit 62 Odd field data storage unit 63 Decimation operation unit 74 Data storage unit

Continuing on the front page (72) Inventor Takahiro Nakayama Aoba-maki Aoba, Aoba-ku, Sendai-shi, Miyagi (No address) Inside Tohoku University (72) Inventor Masahiro Akina Aramaki-Aoba, Aoba-ku, Aoba-ku, Sendai city, Miyagi prefecture (No number) Tohoku university (72) Inventor Hirotsugu Iemura 4-1-1 Hongo, Bunkyo-ku, Tokyo Cosmos Hongo Building I & F Corporation F-term (reference) 5C059 KK11 LC03 MA27 MD07 PP01 SS20 TA57 TB01 TC00 TD10 UA02 UA31 UA39 5J064 AA01 AA02 BA13 BC01 BD03

Claims (37)

[Claims] 1. An original image data in which odd field data and even field data are alternately arranged, a field detecting means for extracting image data of one of the fields, and a field detecting means for extracting the image data of the one field. The image data of the other field is to be compressed, and a data string having at least one or more data is divided into blocks to form a vector, and a code vector similar to a vector extracted from the compression target from a codebook prepared in advance. And a vector quantization means for outputting a code corresponding thereto. 2. The data compression apparatus according to claim 1, further comprising LZSS dictionary compression means for performing LZSS dictionary compression on the code string output from said vector quantization means. 3. A data sequence having at least one or more data is divided into blocks to obtain a vector, and a code vector similar to a vector extracted from a compression target is searched for in a prepared code book, and the corresponding code vector is searched for. A vector quantization unit that outputs a code, and detects a most frequent code from each of the codes output from the vector quantization unit, and detects a code similar to the most frequent code among the output codes. A data compression device comprising: code replacement means for replacing with a specific code. 4. An LZSS for a code string after replacement with a specific code by the code replacement means.
The data compression apparatus according to claim 3, further comprising LZSS compression means for performing dictionary compression. 5. A flag byte header for distinguishing a code replaced by the specific code from a code not replaced and generating a position indicating the position of the code is generated.
4. The data compression apparatus according to claim 3, further comprising a flag byte header conversion unit that deletes a specific code replaced by the code replacement unit from each code output from the vector quantization unit. . 6. The flag byte header conversion means, wherein each of the codes output from the vector quantization means includes:
6. The data compression apparatus according to claim 5, further comprising an optimization unit configured to perform a process of combining the significant bits of each upper byte of the two codes with respect to each code other than the specific code that has not been deleted. apparatus. 7. An LZ which performs LZSS dictionary compression on a data string output from said flag byte header conversion means.
7. The apparatus according to claim 5, further comprising an SS compression unit.
A data compression device according to claim 1. 8. An original image data in which odd field data and even field data are alternately arranged, comprising field detection means for extracting image data of one of the fields, wherein the vector quantization means comprises: 8. The data compression apparatus according to claim 4, wherein vector quantization is performed on image data of one field extracted by said field detection means as a compression target. 9. A data sequence having at least one or more pieces of data is divided into blocks, and a vector is prepared. A code vector similar to a vector extracted from the compression target is searched from a prepared code book, and a corresponding vector is searched for. A data compression apparatus comprising: vector quantization means for outputting a code to be converted; and LZSS compression means for performing LZSS dictionary compression on the code string output from the vector quantization means. 10. The image data to be compressed includes a luminance signal and a chrominance signal. Each pixel value of the chrominance signal is calculated for each of a plurality of pixels, and a color signal of the plurality of pixels is calculated based on the calculated value. The data compression device according to claim 1, further comprising a rate control unit to be represented. 11. A flag byte header for discriminating predetermined specific data from other data and a position of the data based on a data string input as a compression target, and generating the flag byte header. A data compression apparatus comprising: a flag byte header conversion unit that sequentially performs a process of deleting the specific data from each data input as a compression target in units of a predetermined number of data strings. 12. A data sequence having at least one or more data is divided into blocks and taken out from image data to be compressed, and a minimum value of each data in the block is extracted for each block. A data dividing unit that divides the data sequence in the block into a minimum value component and a formed component excluding the minimum value component by subtracting the extracted minimum value from each data, and generating the data sequence for each block by the data dividing unit. Vector quantizing means for searching for a code vector similar to the formed vector from a previously prepared code book and outputting a code corresponding to the code vector prepared in advance, as a vector. Wherein the minimum value component and the formed component are processed separately. . 13. An original image data in which odd field data and even field data are alternately arranged, further comprising field detection means for extracting image data of one of the fields, wherein the data division means comprises: 13. The data compression apparatus according to claim 12, wherein a process of dividing the data of one field extracted by the field detection unit into the formation component and the minimum value component is performed. 14. The image data to be compressed includes a luminance signal and a color signal, calculates each pixel value of the color signal for each of a plurality of pixels, and obtains a color signal of the plurality of pixels by an obtained operation value. The data dividing means performs a process of dividing the data of the luminance signal and the color signal output from the rate controlling means into the formation component and the minimum value component. 14. The data compression device according to claim 12, wherein: 15. A flag byte header for discriminating predetermined specific data from other data and representing the positions of the data based on the code string output from the vector quantization means. 13. The apparatus according to claim 12, further comprising: a flag byte header conversion unit for sequentially performing a process of deleting the specific data from each code output from the vector quantization unit in units of a predetermined number of data strings. 15. The data compression device according to any one of items 14 to 14. 16. A flag byte header converting unit which detects a most frequent code and its appearance frequency from each of the codes output from the vector quantization unit, and based on the detected frequent code appearance frequency. 16. The data compression apparatus according to claim 15, wherein it is determined whether or not to perform the processing. 17. An LZSS dictionary which performs LZSS dictionary compression on a data string output from said flag byte header conversion means.
13. A ZSS compression means is provided.
17. The data compression device according to any one of items 16. 18. The image processing apparatus according to claim 12, further comprising a gradation quantization unit that performs a process of reducing the number of gradations on the data of the minimum value component divided for each block by the data division unit. 18. The data compression device according to any one of claims 17 to 17. 19. A flag byte header for discriminating predetermined specific data from other data and indicating the position of the data based on the data string output from the gradation quantization means. And a flag byte header converting means for sequentially performing a process of deleting the specific data from each data output from the vector quantization means in units of a predetermined number of data strings. 19. The data compression device according to 18. 20. Among the data output from the gradation quantization means, the most frequent data and its appearance frequency are detected, and the flag byte header conversion is performed based on the detected frequent data appearance frequency. 19. The data compression apparatus according to claim 18, wherein it is determined whether or not to perform the processing of the means. 21. After deleting data of one field from original image data in which data of an odd field and data of an even field are alternately arranged, at least one of the data of the other field is set as a compression target. A data sequence having the above data is divided into blocks to form a vector, a code vector similar to the vector extracted from the compression target is searched for from a prepared code book, and the corresponding code is output. A data compression method, characterized in that: 22. The output code sequence is LZ
The data compression method according to claim 21, wherein SS dictionary compression is performed. 23. A data string having at least one or more pieces of data is divided into blocks to obtain a vector, and a code vector similar to a vector extracted from a compression target is searched for from a code book prepared in advance, and a corresponding code vector is searched for. Outputting a code, detecting a most frequent code from the output codes, and replacing a code similar to the most frequent code among the output codes with a specific code. Characteristic data compression method. 24. The data compression method according to claim 23, wherein LZSS dictionary compression is performed on the code sequence after the replacement with the specific code. 25. A flag byte header for distinguishing a code replaced by the specific code from a code not replaced and generating a flag byte header indicating a position of the code, and, among the output codes, 24. The data compression method according to claim 23, wherein the replaced specific code is deleted. 26. The data compression according to claim 25, wherein for each code other than the specific code that has not been deleted, a process of combining valid bits of upper bytes of two codes is performed. Method. 27. An LZSS dictionary compression is performed on the data string after the generation of the flag byte header and the deletion of the specific code, or on the data string after the processing of putting together the valid bits. 27. The data compression method according to claim 25, wherein the method is performed. 28. A data sequence having at least one or more data is divided into a vector, and a code vector similar to the vector extracted from the compression target is searched from a prepared code book,
A data compression method comprising: outputting a code corresponding to the code; and performing LZSS dictionary compression on the output code sequence. 29. The image data to be compressed includes a luminance signal and a chrominance signal, and computes each pixel value of the chrominance signal for each of a plurality of pixels. The data compression method according to any one of claims 21 to 28, characterized in that a process for representing is performed before vector quantization. 30. A data string having at least one or more data items is extracted from image data to be compressed in blocks, and the minimum value of each data item in each block is extracted for each block. By subtracting the extracted minimum value from each data, the in-block data sequence is divided into a minimum value component and a formation component excluding the minimum value component, and the formation component generated for each block by the division is divided. Each data sequence is a vector, a code vector similar to the vector of the formation is searched out from a code book prepared in advance, a code corresponding to the vector is output, and the minimum value component and the formation are separately determined. A data compression method characterized by being processed. 31. The image data to be compressed is formed by alternately arranging odd field data and even field data, and deleting one of the odd field data and the even field data. 31. The data compression method according to claim 30, wherein the data of the other field is divided into the formation component and the minimum value component. 32. The image data to be compressed includes a luminance signal and a color signal. Each pixel value of the color signal is calculated for each of a plurality of pixels, and a color signal of the plurality of pixels is calculated based on the calculated value. 32. The data compression according to claim 30 or 31, wherein after performing a process representative of the above, each data of the luminance signal and the chrominance signal obtained thereby is divided into the formed component and the minimum value component. Method. 33. On the basis of the output code sequence, a flag byte header for distinguishing predetermined specific data from other data and indicating the position of the data is generated, and 33. The data compression method according to claim 30, wherein a process of deleting the specific data from the respective codes is sequentially performed in units of a predetermined number of data strings. 34. Detecting the most frequent code and its appearance frequency from the output codes, generating the flag byte header based on the detected frequent code appearance frequency, and generating the specific data. 34. The data compression method according to claim 33, wherein it is determined whether or not to perform a deletion process. 35. The data compression according to claim 30, wherein LZSS dictionary compression is performed on the data string after the generation of the flag byte header and the deletion of the specific data. Method. 36. A computer-readable storage medium storing a program for causing a computer to function as each unit of the data compression apparatus according to claim 1. Description: 37. A computer-readable storage medium storing a program for causing a computer to execute the procedure of the data compression method according to claim 21.