JP2003318740A - Variable length encoding method and variable length decoding method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve encoding efficiency for variable length encoding in image encoding. <P>SOLUTION: An RL row generating portion 201 makes a frequency coefficient value dimensional by a predetermined scanning order from a low frequency component to a high frequency component, and generates a combined (RL value) row of a successive number R of coefficients whose coefficient value is 0 and a following coefficient value L. A rearranging portion 202 rearranges a row of the RL value in inverse order. A cord assigning portion 203 converts the row of RL value, from the top, into variable length codes by using a first variable length conversion table, and converts the following RL value into the variable length codes by using a second variable length conversion table, when an absolute value of L of the RL value which is a conversion object becomes equal to or more than a predetermined threshold value. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for coding and decoding coefficient values of image data when the image data is coded and decoded by using arithmetic coding.

[0002]

2. Description of the Related Art In a moving image coding process, the amount of information is compressed by utilizing the spatial and temporal redundancy of moving images. Generally, as a method of utilizing the redundancy in the spatial direction, conversion into the frequency domain is used, and as a method of utilizing the redundancy in the time direction, an inter-picture predictive coding process is used.

H. In the moving picture coding method under standardization as 26L, a mode using arithmetic coding has been proposed as a variable length coding method. Here, frequency conversion is performed for each block having a size of 4 × 4 pixels, and further quantization is performed to generate coefficient values. Then, scanning is performed from the DC component toward the high frequency component, and continuous values are 0
Then, a combination of the number R of coefficients and the subsequent coefficient value L is created. Then, these R and L values are binarized into 0 and 1 using a binarization table, and then binary arithmetic coding is performed. For example, a table as shown in FIG. 8 is used as the binarization table for the absolute value of the coefficient value L.

[0004]

SUMMARY OF THE INVENTION In the above conventional method,
When the absolute value of the coefficient value L increases, the code length increases,
This leads to a decrease in coding efficiency. In general, since the absolute value of the coefficient value L of the low frequency component value is a large value, the coding efficiency when coding the low frequency component value is significantly reduced.

[0005]

In order to achieve the above object, the variable length coding method of the present invention has a structure according to claim 1 in which image data is divided into blocks having a predetermined size, A variable-length encoding method for encoding frequency coefficient values of the block, wherein the frequency coefficient values in the block are one-dimensionalized in a predetermined scanning order from a low frequency component to a high frequency component, and a coefficient value A combination of the number R of consecutive coefficients whose value is 0 and the subsequent coefficient value L (RL value)
And a sequence of RL values generated in the combination sequence generation step,
The rearrangement step of rearranging in the reverse order and the sequence of L values rearranged by the rearrangement step are the first L value or the immediately preceding absolute value of L is less than or equal to a predetermined threshold value. In this case, the first binarization table is used for conversion into a binarization sequence, and when the immediately preceding absolute value of L exceeds the threshold value, the second binarization table is used for conversion to 2 It has a configuration including a binarization step of converting into a binarized sequence and an arithmetic coding step of performing arithmetic coding on the binarized sequence obtained by the binarization step and converting into a variable length code. is doing.

According to a sixth aspect of the variable length decoding method of the present invention, the image data is divided into blocks each having a predetermined size, and the frequency coefficient values of the blocks are scanned in a predetermined scanning order. One-dimensional, continuous coefficient value is 0
A variable length decoding method for decoding a code string generated by coding a combination (RL value) of a continuous number R of coefficients and a subsequent coefficient value L. Arithmetic decoding step for decoding the binary sequence into a binary string using arithmetic decoding, and 2 obtained by the arithmetic decoding step.
For the binarized string, if it is the first binarized value or the absolute value of L immediately before is less than or equal to a predetermined threshold value, the absolute binarization of L is performed using the first binarization table. Converting to an absolute value of L by using a second binarization table when the absolute value of L immediately before exceeds the threshold value. A rearrangement step of rearranging the sequence of RL values generated by the multi-value quantization step in reverse order, and an RL rearranged by the rearrangement step.
It has a configuration including a coefficient generation step of returning a sequence of values to the number R of consecutive coefficients having a value of 0 and the subsequent coefficient value L and generating a frequency coefficient value from the combination thereof.

[0007]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described with reference to the drawings. The first embodiment shows a case where intraframe coding processing is performed using the variable length coding method of the present invention. (First Embodiment) FIG. 1 shows a block conversion unit 101,
FIG. 3 is a block diagram of a moving picture coding method according to the present invention, which includes a frequency conversion unit 102, a quantization unit 103, and a variable length coding unit 104. The input image is input to the block conversion unit 101, divided into blocks having a size of 4 horizontal pixels and 4 vertical pixels, and input to the frequency conversion unit 102. The frequency conversion unit 102 performs frequency conversion on the input pixel block and converts it into a frequency coefficient. The converted frequency coefficient is input to the quantization unit 103.

The quantizing unit 103 performs a quantizing process on the input frequency coefficient. Here, the quantization processing means dividing the frequency coefficient value by a predetermined quantization value. Here, the quantized value generally differs for each block and each frequency band. The quantized frequency coefficient is input to the variable length coding unit 104. The structure of the variable length coding unit 104 is shown in FIG. As shown in FIG. 2, the variable length coding unit 104 includes an RL sequence generation unit 201, a rearrangement unit 202, a binarization unit 203, a table storage unit 204, and an arithmetic coding unit 205.

The quantized frequency coefficient output from the quantizer 103 is input to the combination sequence generator 201. In the RL string generation unit 201, the quantized frequency coefficient value is first dimensionalized by a predetermined scanning method. Then, the number R of consecutive coefficient values 0 with respect to the one-dimensionalized coefficient value R
And a subsequent coefficient value L other than 0 (hereinafter,
A sequence of RL values) (hereinafter referred to as RL sequence) is generated. An example thereof will be described with reference to FIGS.

FIG. 3A shows the quantized frequency coefficients output from the quantizer 103. Here, it is assumed that the coefficient at the upper left indicates a DC component, the frequency component in the horizontal direction increases toward the right, and the frequency component in the vertical direction increases toward the bottom. FIG. 3B shows a scanning method for one-dimensionalizing the quantized frequency coefficient values. One-dimensionalization is performed by scanning from the low frequency region to the high frequency region. Then, the result of generating the RL sequence for the one-dimensionalized coefficient value is shown in FIG. Here, in FIG. 4A, the first number indicates the number of coefficients. Generally, the higher the frequency component, the more easily the coefficient value becomes 0. Therefore, by scanning from the low frequency region toward the high frequency region, the amount of information in the RL sequence can be reduced. The generated RL sequence is input to the rearrangement unit 202.

The rearrangement unit 202 rearranges the input RL sequences in the reverse direction. However, the number of coefficients is excluded from the sorting target. The state after rearrangement is shown in FIG.
Becomes The RL string thus rearranged is input to the code assigning unit 203.

In the binarization unit, R,
The value of L is binarized, that is, converted to 0 or 1.
Here, the value of L is assumed to be encoded separately from the absolute value and the sign. The binarization is performed using a predetermined binarization table. The binarization table is held in the table storage unit 204, and the binarization is performed by referring to the table storage unit 204. An example of the binarization table for the absolute value of L is shown in FIG. Since the minimum absolute value of L is 1, the binarization table for the value 0 is not defined as shown in FIG. In the binarization table, generally, the larger the absolute value of L, the longer the binarization value is assigned. Here, although FIG. 5 shows two binarization tables, L
When the absolute value of is small, the binary table 1 is assigned a shorter binary value, and when the absolute value of L is larger, the binary table 2 is assigned a shorter binary value. There is.

For the first absolute value of L, the binarization table 1 is used. For example, the binarized value for the absolute value of the first L value (−2) is “01”. Then, the absolute value of L is binarized in order, and when the absolute value of L exceeds the threshold value, the binarization table 2 is used for the next absolute value of L. When the threshold value for the absolute value of L is 5, the absolute value exceeds the threshold value at the third value (6) of L. Therefore, the first to third absolute values of L are binarized using the binarization table 1.
The fourth absolute value of L is converted using the binarization table 2. Here, at the fifth L value (2), the absolute value of L becomes below the threshold value again, so
The second absolute value of L is binarized again using the binarization table 1. By such binarization processing, 2
The absolute values of the 1st to 5th L values (-2, 3, 6, -12) are binary values "1", "01", "00001", and "00", respectively.
01100 ". Then, the binarized R value,
The absolute value of L and the like are converted into a variable-length code using the arithmetic code in the arithmetic coding unit 205 and output.

As described above, according to the variable length coding method of the present invention, the frequency coefficient values in the block are scanned from the low frequency region to the high frequency region to be one-dimensional. Then, for a one-dimensionalized coefficient value, a sequence of RL values, which is a combination of the number R of consecutive coefficient values 0 and the subsequent coefficient value L other than 0, is generated. Then, the RL value is converted into a variable length code in the reverse order of the scanning order. When converting to a variable length code,
The R value, the absolute value of L, and the sign of L are individually converted. At the time of these conversions, first, binarization is performed. When binarizing the absolute value of L, a plurality of binarizing tables are prepared. Then, first, conversion is performed using the first binarization table, and L
When the absolute value of exceeds the threshold value, the second absolute value of L is converted using the second binarization table. In this case, the first binarization table has a configuration in which the binarization value becomes shorter when the absolute value of L is smaller than that of the second binarization table. The table is configured such that the binarized value becomes shorter when the absolute value of L is larger than that of the first binarized table.

In general, the lower the frequency component, the larger the absolute value of L. Therefore, the RL is scanned by scanning from the low frequency region to the high frequency region.
When the RL value is converted into a variable length code in the reverse order of the value being generated, the absolute value of L gradually increases. Therefore, the total code amount can be reduced by using the binarization table in which the length of the binarized value is shortened when the absolute value of L is large after the absolute value of L exceeds the threshold value. .

In the present embodiment, the case of encoding an image by intra-frame encoding has been described, but this is a case of performing inter-frame encoding on a moving image input using motion compensation or the like. However, the same effect can be obtained by the method of the present embodiment. Further, although the case where the input image is divided into blocks of horizontal 4 pixels and vertical 4 pixels has been described in the present embodiment, the size of the block may be another size.

In the present embodiment, the method of scanning within a block has been described with reference to FIG. 3B, but this is another scanning order if scanning from a low frequency region to a high frequency region. May be. Further, although the present embodiment has been described with reference to FIG. 5 as an example of the binarization table, this may be another table. Further, although the case where two binarization tables are used has been described, this uses three or more binarization tables, uses a plurality of threshold values, and creates a binarization table each time each threshold value is exceeded. You may make it switch. Further, in the present embodiment, the scanning method shown in FIG.
The case where the one-dimensionalized RL sequence is obtained has been described, but this may be another method.

In the present embodiment, when the absolute value of L exceeds the threshold value, the second absolute value of 2 is applied to the next absolute value of L.
If the absolute value of L is less than or equal to the threshold value,
The case where the first binarization table is used for the next absolute value of L has been described, but once the absolute value of L exceeds the threshold value, the absolute value of L thereafter becomes May always use the second binarization table. In this case, in the examples of FIGS. 3 and 4, when the threshold value for the absolute value of L is 5, the absolute value exceeds the threshold value at the third value (6) of L. Therefore, 1-3
The binarization table 1 is used to binarize the absolute value of the th L, and the absolute values of the fourth and subsequent Ls are binarized.
Conversion is performed using the binarization table 2. Here, even if the absolute value of L becomes the threshold value or less again at the fifth L value (2), the binarization table 1 is not used to perform binarization again.

In this embodiment, the case where the rearrangement unit 202 rearranges the RL sequences in the reverse order and performs binarization on the RL values rearranged in the reverse order has been described. The binarization may be performed without rearranging. However, in this case, the binarization unit binarizes the first absolute value of L using the second binarization table of FIG. If the absolute value of L is less than or equal to the threshold value, the first binarization table is used for the next absolute value of L. If the absolute value of L exceeds the threshold value, the next L The second binarization table is used for the absolute value of. Alternatively, the binarization unit binarizes the first absolute value of L using the second binarization table of FIG. Then, when the absolute value of L becomes equal to or less than the threshold value, the first binarization table is used for the subsequent absolute values of L. As a result, the same effect as that of the above embodiment can be obtained. At this time, the rearrangement unit 202 in FIG. 2 is unnecessary.

(Second Embodiment) FIG. 6 shows a variable length decoding unit 601, an inverse quantization unit 602, and an inverse frequency conversion unit 60.
3 is a block diagram of a variable length decoding method of the present invention including a frame memory 604. FIG. Here, it is assumed that the code string generated by the variable length coding method of the present invention described in the first embodiment is input. The code string is input to the variable length decoding unit 601.

The configuration of the variable length decoding unit 601 is shown in FIG. As shown in FIG. 7, the variable length decoding unit 601 includes an arithmetic decoding unit 701, a multi-value quantization unit 702, a table storage unit 703,
It is composed of a rearrangement unit 704 and a coefficient generation unit 705.
The arithmetic decoding unit 701 first performs arithmetic decoding on the input code string. A binary data string is obtained by arithmetic decoding. The binarized data string obtained by arithmetic decoding is output to the multi-value quantization unit 702.

The multi-value conversion unit 702 converts the binarized data string into the R value, the absolute value of L, and the sign of L. Here, the conversion of L into an absolute value will be described. The conversion of L into an absolute value is performed using a plurality of binarization tables. This binarization table is held in the table storage unit 703, and by referring to the table storage unit 703, conversion into multivalued data (absolute value of L) is performed. An example of the binarization table is shown in FIG. Two types of binarization tables are held. Binarization table 1 when the absolute value of L is small
Is assigned a shorter code, and when the absolute value of L is larger, the binarization table 2 is assigned a shorter code. Here, the head part of the binarized data string of the input code string is 01001000001000110.
It is assumed to be 0. 2 for the first binarized data
The binarization table 1 is used. Now, when referring to the binarization table 1 of FIG.
It is the binarization sequence 01, and the absolute value of L in this case is 2.

Similarly, when the conversion from the binarization sequence to the absolute value of L is performed in order, the binarization table 1 is used.
The binarization sequence 001 becomes the absolute value 3 of L and the binarization sequence 00000
1 is converted to the absolute value 6 of L.

Here, when the obtained absolute value of L exceeds a threshold value, the multi-value quantization unit 702 uses the binarization table 2 in the conversion from the next binarization sequence to the absolute value of L. To use. If the threshold value for the absolute value of L is 5, then the third L
The absolute value of 6 will exceed the threshold value. Therefore, the binarization table 2 is used to perform conversion on the fourth binarized column. Therefore, the next variable length code 0001100 is converted into the absolute value 12 of L. Here, when the absolute value of L becomes equal to or less than the threshold value again, the binarization table 1 is used again for the next binarization sequence.
Convert to the absolute value of. As described above, when the absolute value of L, the value of R, and the sign of L for one block are generated, they are input to the rearrangement unit 704. Here, it is assumed that the RL sequence in FIG. 4B is generated.

The rearrangement section 704 rearranges the input RL sequences in the reverse direction. However, the number of coefficients is excluded from the sorting target. The state after rearrangement is shown in FIG.
Becomes The RL sequence thus rearranged is input to the coefficient generation unit 705.

The coefficient generator 705 converts the input RL sequence into coefficient values. At this time, the coefficient value 0 is generated by the value indicated by R based on the predetermined scanning order,
Then, the coefficient value of the value indicated by L is generated repeatedly to convert the RL sequence into the coefficient value. Here, if scanning is performed in a zigzag manner from the low frequency region to the high frequency region, the RL row in FIG.
It will be converted into the coefficient block of (a). The generated coefficient block is input to the inverse quantization unit 602.

The inverse quantization unit 602 performs inverse quantization processing on the input coefficient block. Here, the inverse quantization process means to add a predetermined quantized value to the coefficient value. Here, the quantized value is generally different for each block and each frequency band and is obtained from the code string. The inversely quantized coefficient block is input to the inverse frequency transform unit 603.

The inverse frequency converter 603 performs inverse frequency conversion on the inversely quantized coefficient block to convert it into a pixel block. The converted pixel block is input to the frame memory 604. Decoded pixel blocks are sequentially stored in the frame memory 604, and when one screen of pixel blocks are stored, the decoded pixel blocks are output as an output image.

As described above, in the variable length decoding method of the present invention, when the binary sequence obtained by arithmetically decoding the input code sequence is converted into the absolute value of L, first of all, Convert using a binarization table. When the absolute value of L exceeds the threshold value, the second binarization sequence for the next binarization sequence becomes the second
2 binarization table is used. If the absolute value of L is less than or equal to the threshold value, for the next binarized sequence,
The first binarization table is used. And the obtained RL
After rearranging the values in reverse order, the RL value is converted into a coefficient value based on a predetermined scanning order within the block.

By the above operation, by using the variable length decoding method of the present invention, it is possible to correctly decode the code string encoded by the variable length coding method of the present invention. In the present embodiment, the case of decoding the code string generated by the intra-frame coding has been described, but this is generated by performing inter-frame coding on a moving image input using motion compensation or the like. It is also possible to decode the generated code string, and the method of the present embodiment can obtain the same effect.

Further, in the present embodiment, the case where the input image is divided into blocks of horizontal 4 pixels and vertical 4 pixels and coded is described, but the size of the block is other size. Is also good. In addition, although the present embodiment has been described with reference to FIG. 3B as a scanning method within a block, this may be another scanning order as long as it is the same as the scanning method used at the time of encoding. Is also good.

In the present embodiment, an example of the binarization table has been described with reference to FIG. 5. However, if this is the same as the binarization table used at the time of encoding, another table is used. It may be. Further, although the case where two binarization tables are used has been described, this uses three or more binarization tables, uses a plurality of threshold values, and creates a binarization table each time each threshold value is exceeded. You may make it switch. However, the configuration of the binarization table and the threshold value at this time must be the same as those used at the time of encoding.

Further, in this embodiment, when the absolute value of L exceeds the threshold value, the second absolute value of 2 is applied to the next absolute value of L.
If the absolute value of L is less than or equal to the threshold value,
The case where the first binarization table is used for the next absolute value of L has been described. This is because once the absolute value of L exceeds the threshold value, the subsequent binarization sequence is performed. May always use the second binarization table. In the above embodiment, the absolute value (6) of the third L exceeds the threshold value. Therefore, using the binarization table 1 for the absolute values of the first to third L
The binarization table 2 is used to perform the binarization, and the fourth and subsequent absolute values of L are transformed. here,
Even if the obtained absolute value of L falls below the threshold value again, the binarization table 1 is not used to perform binarization again.

Further, although the case where the RL sequence is encoded in the reverse order to the original order has been described in the present embodiment, when the RL sequence is encoded in the original order,
In the multi-value quantization unit 702, for the first binarization sequence,
Conversion to an absolute value of L is performed using the second binarization table of FIG. Then, if the absolute value of L is less than or equal to the threshold value, the first binarization table is used for the next binarization sequence, and if the absolute value of L exceeds the threshold value, the next 2 The second binarization table is used for the binarization sequence. Alternatively, in the multi-value quantization unit 702, as shown in FIG.
The conversion to the absolute value of L is performed using the second binarization table of Then, when the absolute value of L becomes equal to or less than the threshold value, the first binarization table is used for the subsequent binarization sequence. As a result, the same effect as that of the above embodiment can be obtained. At this time, the rearrangement unit 704 in FIG. 7 is unnecessary.

[0035]

As described above, according to the variable length coding method of the present invention, the frequency coefficient values in a block are scanned from the low frequency region to the high frequency region to be one-dimensional. Then, for a one-dimensionalized coefficient value, a sequence of RL values, which is a combination of the number R of consecutive coefficient values 0 and the subsequent coefficient value L other than 0, is generated. Then, the RL value is converted into a variable length code in the reverse order of the scanning order. When converting to a variable-length code, the R value, the absolute value of L, and the sign of L are individually converted. At the time of these conversions, first, binarization is performed. When binarizing, a plurality of binarizing tables are prepared. Then, first, conversion is performed using the first binarization table, and when the absolute value of L exceeds the threshold value, conversion is performed using the second binarization table for the next absolute value of L. I do. In this case,
The first binarization table is configured such that the binarization value becomes shorter when the absolute value of L is smaller than that of the second binarization table, and the second binarization table is the first binarization table. The binarized value becomes shorter when the absolute value of L is larger than that of the binarized table.

Generally, the lower the frequency component, the larger the absolute value of L. Therefore, scanning from the low frequency region to the high frequency region causes RL to scan.
When the RL value is converted into a variable length code in the reverse order of the value being generated, the absolute value of L gradually increases. Therefore, the total code amount can be reduced by using the binarization table in which the length of the binarized value is shortened when the absolute value of L is large after the absolute value of L exceeds the threshold value. .

As described above, in the variable length decoding method of the present invention, when the binarized string obtained by arithmetically decoding the input code string is converted into the absolute value of L, first of all, Convert using a binarization table. When the absolute value of L exceeds the threshold value, the second binarization sequence for the next binarization sequence becomes the second
2 binarization table is used. If the absolute value of L is less than or equal to the threshold value, for the next binarized sequence,
The first binarization table is used. And the obtained RL
After rearranging the values in reverse order, the RL value is converted into a coefficient value based on a predetermined scanning order within the block. By the above operation, by using the variable length decoding method of the present invention, it becomes possible to correctly decode the code string encoded by the variable length coding method of the present invention.

[Brief description of drawings]

FIG. 1 is a block diagram for explaining an embodiment of the present invention.

FIG. 2 is a block diagram for explaining an embodiment of the present invention.

FIG. 3 is a schematic diagram for explaining an embodiment of the invention.

FIG. 4 is a schematic diagram for explaining an embodiment of the invention.

FIG. 5 is an example of a binarization table for explaining an embodiment of the present invention.

FIG. 6 is a block diagram for explaining an embodiment of the present invention.

FIG. 7 is a block diagram for explaining an embodiment of the present invention.

FIG. 8 is an example of a binarization table for explaining a conventional example.

[Explanation of symbols] 101 Block converter 102 frequency converter 103 quantizer 104 variable length coding unit 201 RL sequence generator 202 Sorting unit 203 Binarization unit 204 table storage 205 arithmetic coding unit

Continued front page    (72) Inventor Makoto Hakai             1006 Kadoma, Kadoma-shi, Osaka Matsushita Electric             Sangyo Co., Ltd. F-term (reference) 5C059 MA05 MA21 MC11 ME01 ME05                       ME11 ME17 NN21 PP04 TA58                       TB08 TC04 TD12 UA02 UA05                       UA38                 5J064 AA02 BA09 BA10 BA16 BB05                       BC01 BC16 BD01

Claims (10)

[Claims] 1. A variable-length coding method when image data is divided into blocks having a predetermined size and the frequency coefficient values of the block are coded, wherein the frequency coefficient value in the block is A combination sequence that generates a sequence of combinations (RL values) of a continuous number R of coefficients having a coefficient value of 0 and a subsequent coefficient value L, which is one-dimensionalized by a predetermined scanning order from low frequency components to high frequency components. A generation step, a rearrangement step of rearranging the sequence of RL values generated by the combination sequence generation step in reverse order, and a sequence of L values rearranged by the rearrangement step with a leading L value. If there is, or if the absolute value of the immediately preceding L is less than or equal to a predetermined threshold value, it is converted into a binary string using the first binarization table, and the absolute value of the immediately preceding L is as described above. When the threshold is exceeded Is a binarization step of converting into a binarization sequence using the second binarization table, and a variable length code by performing arithmetic coding on the binarization sequence obtained by the binarization step. A variable-length coding method, comprising: an arithmetic coding step of converting to. 2. A variable length coding method for dividing image data into blocks having a predetermined size and coding frequency coefficient values of the block, wherein the frequency coefficient values in the block are: A combination sequence that generates a sequence of combinations (RL values) of a continuous number R of coefficients having a coefficient value of 0 and a subsequent coefficient value L, which is one-dimensionalized by a predetermined scanning order from low frequency components to high frequency components. A generation step; a rearrangement step of rearranging the RL value sequence generated by the combination sequence generation step in reverse order; and a L value sequence rearranged by the rearrangement step in order from the top to the first. When the absolute value of L to be converted exceeds a predetermined threshold value by converting into a binary string using the binary table, the second binary table is used for the subsequent values of L. Converted to a binary string Variable-length code, wherein the binary-coding step of performing the binary-coding sequence is performed, and the binary-coded sequence obtained by the binarizing step is subjected to arithmetic coding to convert it into a variable-length code. Method. 3. An arithmetic coding method when image data is divided into blocks having a predetermined size and the frequency coefficient values of the block are coded. Combinatorial column generation for generating a column of combinations (RL values) of a continuous number R of coefficients having a coefficient value of 0 and a subsequent coefficient value L, which is one-dimensionalized by a predetermined scanning order from frequency components to high frequency components If the first value of the L or the absolute value of the immediately preceding L exceeds the predetermined threshold value, the second value Convert to a binary string using the binary table, and if the absolute value of L immediately before is equal to or less than the threshold value, convert it to the binary string using the first binary table. A binarization step, and the binarization step A variable length coding method, comprising: an arithmetic coding step of performing arithmetic coding on the obtained binary string to convert it into a variable length code. 4. An arithmetic coding method when image data is divided into blocks having a predetermined size and the frequency coefficient values of the block are coded, wherein the frequency coefficient value in the block is reduced. Combinatorial column generation for generating a column of combinations (RL values) of a continuous number R of coefficients having a coefficient value of 0 and a subsequent coefficient value L, which is one-dimensionalized by a predetermined scanning order from frequency components to high frequency components And a sequence of L values generated in the combination sequence generation step by using the second binarization table in order from the beginning.
When the absolute value of L to be converted becomes less than or equal to a predetermined threshold value after conversion into a binarized string, the first binarized table is used to convert the subsequent L value into a binarized string. Variable-length code, wherein the binary-coding step of performing the binary-coding sequence is performed, and the binary-coded sequence obtained by the binarizing step is subjected to arithmetic coding to convert it into a variable-length code. Method. 5. The variable length coding method according to claim 1, wherein the first binarization table is compared with the second binarization table. , L is assigned a shorter binarization value when the absolute value is smaller, and the second binarization table is larger than the first binarization table when the absolute value of L is large. The variable-length coding method is characterized in that a longer binarized value is assigned to. 6. The image data is divided into blocks having a predetermined size, the frequency coefficient values of the block are one-dimensionalized in a predetermined scanning order, and the number R of consecutive coefficients having a continuous coefficient value of 0. And a variable length decoding method for decoding a code string generated by coding a string of combinations (RL values) with the following coefficient value L, wherein the code string is subjected to arithmetic decoding. An arithmetic decoding step of decoding into a binary string, and a binary value at the head of the binary string obtained in the arithmetic decoding step, or an absolute value of L immediately before is predetermined. When the absolute value of L is equal to or smaller than the threshold value, the first binarization table is used to perform conversion to the absolute value of L. A multi-value conversion step for converting L into an absolute value using the binarization table of A rearrangement step of rearranging the RL value sequence generated in the multi-value conversion step in reverse order, and the number R of coefficients whose consecutive values are 0 in the RL value sequence rearranged in the rearrangement step and A variable length decoding method, further comprising a coefficient generating step of generating a frequency coefficient value from the subsequent coefficient value L and generating a frequency coefficient value from the combination thereof. 7. The image data is divided into blocks each having a predetermined size, the frequency coefficient values of the blocks are one-dimensionalized in a predetermined scanning order, and the number R of consecutive coefficients having a continuous coefficient value of 0. And a variable length decoding method for decoding a code string generated by coding a string of combinations (RL values) with the following coefficient value L, wherein the code string is subjected to arithmetic decoding. An arithmetic decoding step of decoding into a binary string, and a binary string obtained by the arithmetic decoding step, from the beginning to the absolute value of L using the first binary table. When the conversion is performed and the absolute value of L immediately before exceeds a predetermined threshold value, the second 2 is added to the subsequent binarized value.
A multi-value quantization step of converting the absolute value of L using a binarization table, a rearrangement step of rearranging the columns of RL values generated by the multi-value quantization step in reverse order, A variable characterized in that the sequence of replaced RL values is returned to the number R of coefficients whose consecutive values are 0 and the subsequent coefficient value L, and a coefficient generating step of generating a frequency coefficient value from the combination thereof. Long decoding method. 8. The image data is divided into blocks having a predetermined size, the frequency coefficient values of the block are one-dimensionalized in a predetermined scanning order, and the number R of consecutive coefficients having a continuous coefficient value of 0. And a variable length decoding method for decoding a code string generated by coding a string of combinations (RL values) with the following coefficient value L, wherein the code string is subjected to arithmetic decoding. An arithmetic decoding step of decoding into a binary string, and a binary value at the head of the binary string obtained in the arithmetic decoding step, or an absolute value of L immediately before is predetermined. If the absolute value of L is exceeded by the second binarization table when the absolute value of L is immediately below the threshold value, A multi-value conversion step for converting L into an absolute value using the binarization table of The sequence of RL values generated by the multi-value quantization step is returned to the number R of coefficients whose consecutive values are 0 and the subsequent coefficient value L, and a coefficient generation step of generating a frequency coefficient value from the combination thereof. A variable length decoding method characterized by the above. 9. The image data is divided into blocks each having a predetermined size, the frequency coefficient value of the block is one-dimensionalized in a predetermined scanning order, and the number R of consecutive coefficients having a continuous coefficient value of 0. And a variable length decoding method for decoding a code string generated by coding a string of combinations (RL values) with the following coefficient value L, wherein the code string is subjected to arithmetic decoding. An arithmetic decoding step of decoding into a binary string and a binary string obtained by the arithmetic decoding step are sequentially converted into absolute values of L using a second binary table from the beginning. When the absolute value of L immediately before is converted to a value equal to or smaller than a predetermined threshold value after conversion, the first binarization table is used to convert the subsequent binarized value to the absolute value of L. A multivalued step and a sequence of RL values generated by the multivalued step A variable length decoding method comprising: a coefficient generating step of returning to the number R of coefficients having consecutive values of 0 and the coefficient value L following the coefficient value L, and generating a frequency coefficient value from the combination thereof. 10. The variable length decoding method according to any one of claims 6 to 9, wherein the first binarization table is compared with the second binarization table. , L is assigned a shorter binarization value when the absolute value is smaller, and the second binarization table is larger than the first binarization table when the absolute value of L is large. The variable length decoding method is characterized in that a longer binarized value is assigned to.