JPH0923433A - Method and device for encoding picture - Google Patents

Abstract

PROBLEM TO BE SOLVED: To set the total of code lengths obtained by finally allocating codes to be a minimum. SOLUTION: A code length table generation part 117 generates a table 118 showing the relation of a difference value between motion compensation frames and the code length of the code used in a code allocation part 113 and transmits it to a motion vector detection part 105A. A code length conversion part 202 obtains the code length which is to be allocated to the difference values between the motion compensation frames based on the code length table 118 obtained by a subtracter 201 and inputs it to a total calculation part 203. The total calculation part 203 calculates the total of the code lengths in a small block. A minimum value memory 204 compares the calculated total with a minimum value accumulated in a memory at that time. When the total is smaller than the minimum value, the minimum value is updated and an updated indication instruction is transmitted to a vector memory 206.

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to an image coding method and apparatus.

[0002]

2. Description of the Related Art There is lossless coding that can completely restore an original image, such as broadcasting materials and medical care. On the other hand, motion-compensated prediction is a technique essential for high-efficiency coding of moving images. Motion-compensated prediction makes use of the similarity between a frame to be coded and other frames that are temporally close to each other, and is an effective means for improving coding efficiency even for lossless coding. .

FIG. 4 is a block diagram of a lossless coding apparatus constructed using conventional motion compensation. First, the encoding target image 102 input from the input terminal 101 is divided into N × N encoding target small blocks 104 by the small block dividing unit 103. For each of the divided small blocks 104, the motion vector detection unit 105
The motion vector 108 is selected from the reference image 107 stored in 06, that is, the images that are temporally close to the encoding target image 102, based on a predetermined evaluation criterion. The motion compensation unit 109 obtains the reference small block 110 corresponding to the selected motion vector 108, and the subtractor 111 calculates the difference from the coding target small block 104 to obtain the motion compensation inter-frame difference value 112. After that, the code is assigned by the code assigning unit 113, multiplexed by the multiplexing unit 114 with the motion vector information 108, and then output as the encoded data 115 to the output terminal 116.

Here, in selecting the motion vector, the following method is usually used. Let the image data of the small block 104 to be encoded be N × N in size and B
Let (i, j). With respect to B (i, j), a small block B '(i + u, which spatially and horizontally shifts B (i, j) by (u, v) on the reference image 107 in the frame memory 106. j + v) evaluation value D indicating the degree of similarity
Calculate (u, v). The sum of absolute differences and the sum of squared differences are often used as the evaluation value. For example, in the case of the sum of absolute differences, the calculation is as follows.

[0005]

[Equation 1] By changing (u, v) in the defined search area, D (u,
(u, v) that gives the minimum value of v) is the final motion vector (U, V), and the small blocks in the reference image 107 defined by the motion vector (U, V) are the reference small blocks 110. To be done.

The structure of the above motion vector detecting section 105C will be described in detail. The candidate vector setting unit 205 has a function of generating a vector within the defined search range and transmitting it to the vector memory 206. Shift unit 207
Is the reference image 1 stored in the frame memory 106.
07 is spatially shifted by that amount by the candidate vector set by the candidate vector setting unit 205. Subtractor 2
In 01, the difference between the shifted small block and the encoding target small block 104 is calculated, and the calculation result is input to the absolute value sum calculation unit 214. In the absolute value sum calculation unit 214, the sum of the absolute differences between the small blocks is calculated, and the calculated absolute value sum is sent to the minimum value memory 204 and the minimum value stored in the memory 204 at that time. A comparison is made. If the sum of absolute values is smaller than the minimum value, the minimum value is updated, and an update instruction command is sent to the vector memory 206, and the stored vector is replaced with the candidate vector input at that time. To be done. When the generation of all the vectors in the search range defined by the candidate vector setting unit 205 is completed, the vector stored in the vector memory 206 is output as the motion vector 108 of the encoding target small block 104, and The minimum value memory 204 and the vector memory 206 are reset.

[0007]

According to the conventional method described above, a reference small block having a high degree of similarity with the small block to be encoded can be selected, and the difference value between motion compensation frames becomes small. By designing a suitable variable-length code and assigning the code, efficient coding becomes possible. In general, the variable length code is assigned a short code for a motion-compensated inter-frame difference value having a high occurrence probability and a long code for a low-probability motion compensation frame difference value to shorten the total code length. Is set as follows.

However, in the conventional reference small block selection method, only the relationship between the small block to be coded and the small block of the reference image is used as the evaluation value, and it is completely unclear what code is assigned in the subsequent stage. I'm not aware. Therefore, when the property or compression rate of the input image changes, there is no guarantee that the reference small block will minimize the final code amount.

For example, in the difference value between motion compensation frames,
Assume that a variable length code as shown in Table 1 is applied.

[0010]

[Table 1] In this case, comparing the small block having the motion-compensated inter-frame difference value of FIG. 5 (a) with the small block of FIG. 5 (b), FIG. 5 (b) is better than the conventional selection method based on the sum of absolute differences. It is selected to be smaller, but in this case, the length when the variable length code is actually assigned to the motion compensation interframe difference value is longer in FIG. 5B. in this way,
In the conventional method, since the reference small block for motion compensation is determined without considering the length of the code applied to the motion compensation inter-frame difference value, the finally obtained code length is There was a problem that it was not the shortest.

An object of the present invention is to provide an image coding method and apparatus capable of minimizing the total sum of code lengths obtained by finally allocating codes.

[0012]

According to the image coding method of claim 1, the image data of the current frame is divided into small blocks for the digital image signal, and the divided small blocks to be coded are divided. Obtain a motion vector from past or future reference frames stored in the frame memory,
An image coding method in which a motion compensation interframe difference value between a reference block determined by the motion vector and an encoding target small block of the current frame is calculated, and then coded by assigning a code to the interframe difference value for transmission / recording. In, in each reference block corresponding to a plurality of candidate vectors in the reference frame, the code length for the motion compensation interframe difference value is summed at each pixel in the block, and the sum of the code lengths is minimized. The feature is that the thing is a motion vector.

An image coding method according to a second aspect is the image coding method according to the first aspect, in which a plurality of code tables are adaptively switched to perform coding, and a plurality of candidate vectors in a reference frame are included. For each of the above, the code lengths for the motion-compensated inter-frame difference values are summed at each pixel in the block for each code table, and the one having the smallest sum of the code lengths is set as the motion vector and selected. Code allocation is performed using a code table that gives the specified code length.

According to the image coding method of the third aspect, when the motion vector search is performed by the tree search of N stages, K− from the first stage.
For a one-step search, the absolute value of the motion compensation interframe difference value is summed for each pixel in the block for each of the plurality of candidate vectors in the reference frame, and the sum of the absolute value sums is the smallest. Is the center of the next stage motion vector search,
Regarding the search from the K stage to the N stage, claim 1 or claim 2
Perform a search according to.

According to another aspect of the image coding apparatus of the present invention, a small block dividing unit that divides the input encoding target image into encoding target small blocks, a frame memory that stores a reference image,
A motion vector detection unit that obtains a motion vector from past or future reference images stored in the frame memory, a motion compensation unit that obtains a reference small block corresponding to the obtained motion vector, the reference small block, and the code. A subtractor for calculating a motion-compensated inter-frame difference value that is a difference between the motion-compensated inter-frame difference value, a code allocation unit that allocates a code to the motion-compensated inter-frame difference value, and a motion-compensated inter-frame difference value to which a code is allocated. A multiplexing unit that multiplexes with the motion vector and outputs as encoded data, and creates a code length table that indicates the relationship between the motion compensation interframe difference value and the code length of the code used in the code allocation unit, A code length table creating unit for outputting to the motion vector detecting unit, wherein the motion vector detecting unit is a vector memory; Generate a vector in the search range defined by Kuttle, a candidate vector setting unit for storing in the vector memory, and a reference image stored in the frame memory by the candidate vector generated by the candidate vector setting unit A shift unit that shifts by an amount, a subtractor that calculates a difference value between motion-compensated frames of the shifted small block and the small block to be encoded, and a difference value between the motion-compensated frame difference values based on the code length table. A code length conversion unit that obtains a code length to be assigned to the code block, a total sum calculation unit that calculates the total sum of the code lengths in a small block, and the total sum are compared with a currently stored minimum value, and the total sum is calculated. Is smaller than the minimum value, the minimum value is updated by the sum and a minimum value memory for sending an update instruction command to the vector memory is included. The tor memory replaces the stored vector with the candidate vector input at that time by the update instruction command, and when the generation of all the vectors in the search range defined by the candidate vector setting unit is completed, The stored vector is output as the motion vector of the encoding target small block.

According to another aspect of the image coding apparatus of the present invention, a small block dividing unit that divides the input encoding target image into encoding target small blocks, a frame memory that stores a reference image,
A motion vector detection unit that obtains a motion vector from past or future reference images stored in the frame memory, a motion compensation unit that obtains a reference small block corresponding to the obtained motion vector, the reference small block, and the code. A subtracter that calculates a difference value between motion-compensated frames, which is a difference from the small block to be converted, and a first, second, ..., Nth (n ≧ n) that assigns a code to the difference value between motion-compensated frames.
2) a code assigning unit, a multiplexing unit that multiplexes a motion compensation interframe difference value to which a code is assigned with the motion vector and outputs the coded data, the motion compensation interframe difference value, and the first, The first, second, ..., Nth code length tables showing the relationship with the code lengths of the codes used in the second, ..., Nth code allocation unit are created,
The first, second, ...
.., a first code selection switch that connects the output of the n-th code length table to one of the inputs of the first to n-th code allocation units, and the first to n-th A second code selection switch for outputting any one of the outputs of the code assigning unit to the multiplexing unit, and the motion vector detecting unit stores a vector memory and a vector in the search range in which the motion vector is determined. A candidate vector setting unit that is generated and stored in the vector memory; a shift unit that shifts the reference image stored in the frame memory by the candidate vector generated by the candidate vector setting unit; A subtracter that calculates a motion-compensated inter-frame difference value between the small block and the encoding target small block, and each of the motions based on the first, second, ..., Nth code length tables. In the small block, first, second, ..., Nth code length conversion units for determining the first, second, ..., Nth code lengths to be assigned to the compensation frame difference value , The n-th summation calculation unit for calculating the summation of the first, second, ..., Nth code lengths, and the summation of the first codelengths, The comparison unit that compares the sum of the second code lengths, ..., The sum of the nth code lengths and outputs the smallest sum and the code flag that gives the sum, and the minimum sum are currently stored. If the sum is smaller than the minimum value, the minimum value is updated by the sum and the minimum value memory for sending an update instruction command to the vector memory and the code flag are stored. , The first and second code selection switches so as to select a code allocation unit corresponding to the code flag. A code selection flag memory for switching a switch, and the vector memory replaces a stored vector with a candidate vector input at that time by an update instruction command, and a search range determined by the candidate vector setting unit. When the generation of all the vectors is finished, the stored vector is output as the motion vector of the encoding target small block.

[0017]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention is characterized in that, when searching a motion vector and selecting a reference small block, the length of a code assigned to a motion compensation interframe difference value is taken into consideration. is there.

Specifically, the difference value between the small block to be coded and the small block of the reference image is replaced with the code length of the code used in the subsequent stage as an evaluation value for obtaining the motion vector, and The one having the smallest sum of code lengths for all pixels is set as the final motion vector, and the small block corresponding to the motion vector is set as the reference small block.

First, an N × N small block B to be coded
A small block B ′ (i + u, j + v) obtained by spatially shifting B (i, j) by (u, v) in the horizontal / vertical direction on the reference image in the frame memory with respect to (i, j). The motion compensation frame difference value E _{u, v} (i, j) of is calculated based on the following equation.

[0020]

[Equation 2] Let T [E _{u, v} (i, j)] be the length of the code to be assigned to the difference value E _{u, v} (i, j), and let each vector (u,
For v), calculate L (u, v) in the following equation. T means a conversion table of the difference value and the code length.

[0021]

(Equation 3) By changing (u, v) within the set search area, (u, v) that gives the minimum value of L (u, v) is set as the final motion vector (U, V), and the motion is calculated. Vector (U,
The small block in the reference image defined by V) is set as the reference small block.

Although the above configuration is for the case where only one type of code is used in the subsequent stage, the same method can be used when a plurality of codes are switched and used as in claim 2. In the case of switching and using k codes, there are k kinds of code allocation units, and one of them is selected and coded. In this case, since there are k types of code length tables, the code length conversion unit calculates k total code lengths for one candidate vector, and records the smallest sum in the minimum value memory. In this case, together with the vector giving the minimum value, the code allocating unit is informed of which of the k kinds of codes is used, and the code allocating unit performs encoding with the informed code.

Further, claim 3 is a process when the motion vector search is performed by a tree search. The tree search is shown in Figure 3.
As shown in (1), first, in the first stage, a vector included in the search range is roughly sampled and searched, and in the next stage, a vector in the vicinity of the vector found in the first stage is searched. A final vector is obtained by repeating the same processing N times. In such a tree search, the influence of the characteristics of the code used in the search of the first stage is considered to be small. Therefore, for the search from the first stage to K-1 stages, there are a plurality of candidates in the reference frame. For each of the vectors, the absolute values of the motion-compensated inter-frame difference values are summed at each pixel in the block, and the one having the minimum sum of the absolute values is the center of the motion vector search of the next stage. By adopting such a conventional method, the search according to claim 1 or claim 2 is performed for the search from K stages to N stages where the influence of the actually assigned code length becomes large. By doing this, the operation of mapping the table becomes unnecessary in the search of the first stage,
This leads to faster processing time. Further, referring to FIG. 3, first, a search range of a vector represented by 15 × 15 dots is shown. As the search vector in the first stage, nine vectors marked with a circle are selected as candidates. Next, it is assumed that an evaluation calculation is performed on each candidate vector and, as a result, the vector ◯ in the upper right corner is selected. Next, in the upper right corner
Eight Δ-mark candidate vectors are selected around the mark vector and the same evaluation operation is performed. This time, the Δ-mark vector in the lower right corner is selected. The search is repeated 3 times in total.

According to the method described above, the difference value between the small block to be encoded and the small block of the reference image is replaced with the code length of the code used in the subsequent stage as the evaluation value when the motion vector is obtained. The motion vector having the smallest sum of code lengths for all pixels in the small block becomes the final motion vector, and the small block corresponding to the motion vector becomes the reference small block. As a result, the operation is performed so that the total sum of code lengths finally obtained by assigning codes becomes the shortest.

[0025]

Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of an image coding apparatus according to an embodiment of the present invention. The same symbols as those in FIG. 4 indicate the same things.

In the present embodiment, a code length table creating unit for creating a code length table 118 showing the relationship between the motion compensation interframe difference value and the code length used in the code assigning unit 113 is added to the configuration of FIG. 117 is added, and the motion vector detection unit 105A has the motion vector detection unit 105 in FIG.
The configuration is partially different from C. Motion vector detection unit 10
The code length conversion unit 20 calculates a code length to be assigned to each motion compensation interframe difference value based on the code length table 118 instead of the absolute value sum calculation unit 214.
2 and a total sum calculation unit 203 for calculating the total sum of code lengths in a small block.

Next, the operation of this embodiment will be described.

First, the coding target image 102 input from the input terminal 101 is N × in the small block dividing unit 103.
It is divided into N encoding target small blocks 104. For each of the divided small blocks 104, the motion vector detection unit 10
In 5A, the motion vector 108 is selected based on a predetermined evaluation criterion from the reference image 107 stored in the frame memory 106, that is, the images that are temporally close to the encoding target image 102. In the motion compensation unit 109, the reference small block 110 corresponding to the selected motion vector 108.
Is calculated, the difference from the small block 104 to be encoded is calculated in the subtractor 111, and the motion compensation inter-frame difference value 112 is obtained. Then, the code allocation unit 113 allocates the code, and the multiplexing unit 114 In the above, after being multiplexed with the motion vector information 108, it is output to the output terminal 116 as encoded data 115. The code length table creation unit 117 creates a table 118 showing the relationship between the motion compensation interframe difference value and the code length of the code used in the code allocation unit 113, and the motion vector detection unit 105A.
Sent to

The details of the operation of the motion vector detecting section 105A are as follows. The candidate vector setting unit 205 generates a vector within the defined search range and sends it to the vector memory 206. In the shift unit 207,
Reference image 107 stored in frame memory 106
Is spatially shifted by that amount by the candidate vector set by the candidate vector setting unit 205. Subtractor 20
In No. 1, the motion compensation interframe difference value between the shifted small block and the encoding target small block 104 is calculated and input to the code length conversion unit 202. Code length conversion unit 2
In 02, the code length to be assigned to each motion compensation inter-frame difference value is obtained based on the code length table 118, and then input to the sum calculation unit 203. The sum total calculation unit 203 calculates the sum of the code lengths in the small blocks, sends the calculated sum to the minimum value memory 204, and compares it with the minimum value stored in the memory at that time. If the total sum is smaller than the minimum value, the minimum value is updated, and an update instruction command is sent to the vector memory 206 to replace the stored vector with the candidate vector input at that time. . When the generation of all the vectors in the search range defined by the candidate vector setting unit 205 is completed, the vector memory 206
Is stored as the motion vector 108 of the small block to be encoded, and the minimum value memory 204 and the vector memory 206 are reset.

FIG. 2 is a block diagram of an image coding apparatus according to another embodiment of the present invention.

This embodiment is an example of the case where there are two types of variable length codes. Therefore, two types of code assigning units, that is, the first code assigning unit 113A and the second code assigning unit 113B, and two types of code length table creating units, that is, the motion compensation interframe difference value and the first code assigning unit. 1
A first code length table creation unit 1 that creates a table 118A showing the relationship between the code length and the code length used in 13A.
17A, and a second code length table creation unit 117B that creates a table 118B indicating the relationship between the motion compensation inter-frame difference value and the code length of the code used in the second code allocation unit 113B. Further, in the motion vector detection unit 105B, the first code length that determines the code length to be assigned to each motion compensation interframe difference value based on the first code length table 118A and the second code length table 118B, respectively. The conversion unit 202A, the second code length conversion unit 202B, and the first code length conversion unit 2 respectively.
02A, the code length obtained by the second code length conversion unit 202B is input, and the first sum total calculation unit 203A and the second total sum calculation unit 203B that obtain the sums 208A and 208B of the code lengths are input.
, The code selection flag memory 212, and the sum of code lengths 20
8A, 208B are input and compared, and the smaller sum 211
Is stored in the minimum value memory 204, and the sum of the selected one is 2
The comparison unit 209 for writing the code flag 210 giving 11 to the code selection flag memory 212 is provided.

The code selection switches 119 and 120 are code flags 21 stored in the code selection flag memory 212.
Switch according to 3.

Next, the operation of this embodiment will be described.

First, the coding target image 102 input from the input terminal 101 is divided into 16 by the small block dividing unit 103.
It is divided into × 16 encoding target small blocks 104. For each of the divided small blocks 104, the motion vector detection unit 105B uses the reference image 107 stored in the frame memory 106, that is, the images that are temporally close to the encoding target image 102, according to a predetermined evaluation criterion. The motion vector 108 is selected. In the motion compensation unit 109, the reference small block 1 corresponding to the selected motion vector 108
10 is calculated, the difference from the small block 104 to be encoded is calculated in the subtractor 111, and the motion compensation interframe difference value 112 is obtained, and then the first code allocation unit 11
Codes are assigned by the 3A or second code assigning unit 113B, multiplexed by the multiplexing unit 114 with the motion vector information 108, and then output to the output terminal 116 as encoded data 115.

On the other hand, the first code length table creating section 117
In A, a table 118A showing the relationship between the difference value between motion compensation frames and the code length of the code used in the first code allocation unit 113A is created, and in the second code length table creation unit 117B, the motion compensation frame is created. Difference value and second
A table 118B indicating the relationship with the code length of the code used in the code allocation unit 113B is created and sent to the motion vector detection unit 105B.

The details of the operation of the motion vector detecting section 105B are as follows. Candidate vector setting unit 20
5 generates a vector within the defined search range and sends it to the vector memory 206. Shift unit 20
7, the reference image 107 stored in the frame memory 106 is spatially shifted by that amount by the candidate vector set by the candidate vector setting unit 205. In the subtractor 201, the motion compensation interframe difference value between the shifted small block and the encoding target small block 104 is calculated and input to the first code length conversion unit 202A and the second code length conversion unit 202B. . In the first code length conversion unit 202A, after the code length to be assigned to each motion compensation interframe difference value is obtained based on the first code length table 118A, the first total sum calculation unit 203A
Is input to The first total sum calculation unit 203A calculates the total sum 208A of the code lengths in the small block. Similarly, in the second code length conversion unit 202B, after the code length to be assigned to each motion compensation interframe difference value is obtained based on the second code length table 118B, the second total sum calculation is performed. It is input to the section 203B. In the second total sum calculation unit 203B, the pixel 208B having the code length in the small block is
Is calculated. The calculated two types of sums 208A and 208B are sent to the comparison unit 209, and the smaller sum 21
1 is sent to the minimum value memory 204, and comparison with the minimum value stored in the memory at that time is performed. If the selected sum total 211 is smaller than the minimum value stored in the minimum value memory 204, the minimum value is updated and an update instruction command is sent to the vector memory 206 and stored. Existing vector is replaced by the candidate vector input at that time. Furthermore, the comparator 20
The code flag 210 which gives the total sum 211 of the selected one from 9 is written in the code selection flag memory 212.

When the generation of all the vectors within the search range defined by the candidate vector setting unit 205 is completed,
The vector stored in the vector memory 206 is output as the motion vector 108 of the small block to be encoded, and the code flag stored in the code selection flag memory 212 is output as the code selection flag 213 to the code selection switches 119 and 120. At the same time, the minimum value memory 204, the vector memory 206, and the code selection flag memory 212 are reset. Code selection switch 119,
Reference numeral 120 switches the switch based on the code selection flag 213.

In this embodiment, the following calculation is performed. For a 16 × 16 small block B (i, j) to be encoded, B (i, j) is spatially and horizontally (u, v) only on the reference image 107 in the frame memory 106. The shifted small blocks B '(i + u,
j + v) and a motion compensation inter-frame difference value E _{u, v} (i, j)
Is calculated based on the following.

[0040]

(Equation 4) The lengths of two types of codes to be assigned to the difference value E _{u, v} (i, j) are W ₁ [E _{u, v} (i, j)], W, respectively.
₂ [E _{u, v} (i, j)], L ₁ (u, v) and L ₂ (u, v) in the following equations are calculated for each vector (u, v).

[0041]

(Equation 5) By changing (u, v) within the set search area, the final motion of (u, v) that gives the minimum value of L ₁ (u, v) and L ₂ (u, v) is performed. A reference image 10 defined as a motion vector (U, V) with a vector (U, V)
The small blocks in 7 are referred to as reference small blocks. Further, the code used in the code allocation units 113A and 113B is determined depending on whether the minimum value is given by using W ₁ or W ₂ .

[0042]

As described above, according to the present invention, the difference value between the small block to be encoded and the small block of the reference image is used as the evaluation value when the motion vector is obtained, and the difference is the code length of the code used in the subsequent stage. Substituting, the smallest sum of code lengths for all pixels in the small block is set as the final motion vector, and the small block corresponding to the motion vector becomes the reference small block.
In a lossless coding method in which a code is assigned to a motion-compensated inter-frame prediction difference value and transmitted / recorded, the sum of code lengths finally obtained by assigning codes can be minimized. Further, since the code lengths of the codes to be used are stored in a table, the processing is easy, and it can be said that this is a very highly feasible method.

[Brief description of drawings]

FIG. 1 is a block diagram of an encoding device according to an embodiment of the present invention.

FIG. 2 is a block diagram of an encoding device according to another embodiment of the present invention.

FIG. 3 is a diagram showing a motion vector detection method by tree search.

FIG. 4 is a block diagram of a conventional example of an encoding device.

FIG. 5 is a diagram showing a problem of a conventional encoding method.

[Explanation of symbols]

 101 Input Terminal 102 Coding Target Image 103 Small Block Dividing Unit 104 Coding Target Small Block 105A, 105B Motion Vector Detection Unit 106 Frame Memory 107 Reference Image 108 Motion Vector 109 Motion Compensation Unit 110 Reference Small Block 111 Subtractor 112 Motion Compensation Frame Inter-difference value 113 Code assignment unit 114 Multiplexing unit 115 Coded data 116 Output terminal 117 Code length table creation unit 117A First code length table creation unit 117B Second code length table creation unit 118 Code length table 118A First Code length table 118B Second code length table 119, 120 Code selection switch 201 Subtractor 202 Code length conversion unit 202A First code length conversion unit 202B Second code length conversion unit 203A First sum calculation unit 203B 2 total sum calculation unit 204 minimum value memory 205 candidate vector setting unit 206 vector memory 207 shift unit 208A first code length summation 208B second code length summation 209 comparison unit 210 code flag 211 selected summation 212 code selection Flag memory 213 Code selection flag

Claims (5)

[Claims] 1. For a digital image signal, image data of the current frame is divided into small blocks, and the divided small blocks to be encoded are extracted from past or future reference frames stored in a frame memory. After obtaining a motion vector, calculating a motion compensation interframe difference value between a reference block determined by the motion vector and an encoding target small block of the current frame, and encoding and transmitting / recording by assigning a code to the interframe difference value. In the image coding method described above, for each reference block corresponding to a plurality of candidate vectors in the reference frame, the code lengths for the motion compensation inter-frame difference value are summed at each pixel in the block, and the code length An image coding method characterized in that a motion vector having a minimum sum is used as a motion vector. 2. An image coding method according to claim 1, wherein a plurality of code tables are adaptively switched for coding, wherein motion compensation is performed for each of a plurality of candidate vectors in a reference frame. The code length for the inter-frame difference value is summed at each pixel in the block for each code table, and the one having the smallest sum of the code lengths is set as the motion vector, and the selected code length is given. An image coding method characterized by performing code assignment in a code table. 3. When motion vectors are searched by N-th tree search, the motion-compensated interframe difference for each of a plurality of candidate vectors in the reference frame in the search from the first to K-1th stages. The absolute values of the values are summed at each pixel in the block, and the one having the minimum sum of absolute values is set as the center of the motion vector search in the next stage, and K to N stages are searched. An image coding method characterized by performing a search according to 2. 4. A small block dividing unit that divides an input encoding target image into encoding target small blocks, a frame memory that stores a reference image, and a past or future reference image that is stored in the frame memory. A motion vector detection unit that obtains a motion vector, a motion compensation unit that obtains a reference small block corresponding to the obtained motion vector, and a motion compensation interframe difference value that is a difference between the reference small block and the encoding target small block. , A code assigning unit that assigns a code to the motion compensation inter-frame difference value, a multiplexing unit that multiplexes the motion compensation inter-frame difference value to which the code is assigned with the motion vector, and outputs the coded data. Section and a code length table indicating the relationship between the motion compensation inter-frame difference value and the code length of the code used in the code allocation section. A code length table creating unit for creating a block and outputting it to the motion vector detecting unit. The motion vector detecting unit generates a vector memory and a vector within a search range in which the motion vector is defined. A candidate vector setting unit to be stored in the memory, a shift unit for shifting the reference image stored in the frame memory by the candidate vector generated by the candidate vector setting unit, a shifted small block and the code. A subtractor for calculating a motion compensation inter-frame difference value of a small block to be coded, a code length conversion unit for obtaining a code length to be assigned to each motion compensation inter frame difference value based on the code length table, A sum total calculation unit that calculates the sum total of the code lengths in the block, and the sum total is compared with a currently stored minimum value, and the sum total is calculated. Is smaller than the minimum value, the minimum value is updated by the sum and a minimum value memory for sending an update instruction command to the vector memory is included, and the vector memory stores the vector stored by the update instruction command. The candidate vector input at that time is replaced, and when the generation of all the vectors in the search range defined by the candidate vector setting unit is completed, the stored vector is moved to the small block to be encoded. An image encoding device that outputs as a vector. 5. A small block division unit that divides an input encoding target image into encoding target small blocks, a frame memory that stores a reference image, and a past or future reference image that is stored in the frame memory. A motion vector detection unit that obtains a motion vector, a motion compensation unit that obtains a reference small block corresponding to the obtained motion vector, and a motion compensation interframe difference value that is a difference between the reference small block and the encoding target small block. , A first, second, ..., Nth (n ≧ 2) code assigning unit that assigns a code to the inter-motion-compensation-frame difference value, and a motion-compensated frame to which a code is assigned. A multiplexing unit that multiplexes the difference value with the motion vector and outputs the coded data, the motion compensation inter-frame difference value, and the first, second, ..., Nth code. The first, second, ..., Nth code length tables showing the relations with the code lengths of the codes used in the code assignment unit are created and output to the motion vector detection unit. ..., the n-th code length table creation unit, the first code selection switch that connects the output of the subtractor to one of the inputs of the first to n-th code allocation units, and the first to first code selection switches. A second code selection switch for outputting any output of the n-th code allocation unit to the multiplexing unit, wherein the motion vector detection unit includes a vector memory and a search range in which a motion vector is defined. A candidate vector setting unit that generates a vector and stores it in the vector memory, a shift unit that shifts the reference image stored in the frame memory by the candidate vector generated by the candidate vector setting unit, and a shifter. , The n-th sub-block for calculating a motion compensation inter-frame difference value between the encoded small block and the encoding target small block, and the first, second, ...
, Which should be assigned to each motion compensation inter-frame difference value based on the code length table of
The first, second, ..., Nth code length conversion units for obtaining the nth code length, and the first, second, ...
.., first, second, ... for calculating the sum of the nth code lengths
.., the sum of the first code length, the sum of the second code length, ..., The sum of the n-th code lengths are compared, and the smallest sum and the sum are calculated. A comparison unit that outputs a given code flag and the minimum total sum are compared with the currently stored minimum value. If the total sum is smaller than the minimum value, the minimum value is updated with the total sum, and A minimum value memory that sends an update instruction command to a vector memory, and a code selection flag that switches the first and second code selection switches so as to store the code flag and select a code allocation unit corresponding to the code flag. The vector memory includes a memory and replaces the stored vector with a candidate vector input at that time by an update instruction command, and all vectors in the search range defined by the candidate vector setting unit. When the occurrence of is completed, the image encoding device that outputs are accumulated vector as the motion vector of the encoding target small block.