JPH0998433A - Device and method for transmitting image signal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce an encoded data amount to be transmitted by adaptively significance-judging motion compensation residual signals and to control the encoded data amount to be transmitted to be fixed at all times by performing feedback/feedforward buffering. SOLUTION: As for input image data, motion vector is detected by a block unit by block matching for instance with the locally decoded images of a frames in the past. By the motion vector of the block unit, motion compensation is performed to a previous frame, predicted pictures are prepared and a predicted residual with the input image data is calculated. A target bit rate is reset from a bit rate and the buffer residual amount of a framing circuit 17 serving also as a buffer, and for the predicted residual, respective hierarchical branches are predictively calculated beforehand and the hierarchical branch to be the reset target bit rate is selected. Significance judgement is performed for the hierarchical branch, quantization is performed, then variable length encoding is performed and the encoded data mount is transmitted through the framing circuit 17 serving also as the buffer.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image signal transmitting apparatus and method capable of further reducing the data transmission amount when compressing a blocked image signal.

[0002]

2. Description of the Related Art Conventionally, in image compression, a general method is to reduce the amount of information by quantizing a prediction residual obtained by some preprocessing with a predetermined number of bits. As one of effective methods for reducing the data amount (hereinafter, referred to as the encoded data amount) generated by quantizing the prediction residual, the difference between the supplied sample value and its predicted value is calculated. By doing so, the obtained prediction error is quantized with a predetermined number of bits, but since all prediction residuals are quantized, the amount of encoded data becomes considerable.

As one method of suppressing the amount of coded data, the prediction residual is divided into blocks, and a significance judgment is made to judge whether or not it is significant to transmit each block.
In the significance judgment, there is a method in which a block judged to be insignificant is not transmitted.

[0004]

However, in this case, the probability of appearance of a block that becomes significant when the block size is small is reduced, but since the total number of blocks is increased, overhead such as flags is increased, resulting in encoding. There was a problem that the amount of data would increase.
In addition, if the block size is small, the number of significant blocks changes significantly depending on the content of the image, which makes it difficult to control the amount of information to be constant.

Therefore, an object of the present invention is to reduce the amount of encoded data by using a hierarchical structure and to control the amount of encoded data below a target value by performing feedback / feedforward buffering. An object of the present invention is to provide an image signal transmission device and method capable of performing the image signal transmission.

[0006]

According to a first aspect of the present invention, there is provided a video signal transmission apparatus for encoding a prediction residual between an input pixel value and its prediction value, wherein 1 block, the first block is subdivided, and a first block significant judgment means for judging whether or not the first block is significant based on the activity of the first block, The second block significant determination means for generating the second block and performing the significant determination for each second block, the first block significant determination means, and the second block significant determination means are significant. The quantizing means for quantizing the prediction residual included in the determined block, the buffer means for receiving the output of the quantizing means and outputting the output of the quantizing means at a constant rate, and the remaining buffer capacity of the buffer means When there are many , The target value is set based on the remaining buffer so as to increase the target value, and the amount of information to be transmitted is set by predicting the amount of information generated by varying the threshold value for significance judgment. Less than or equal to the value,
An image signal control means configured to determine a threshold value and supply the determined threshold value to the significance determining means for the first and second blocks. It is a transmission device.

Further, in the invention according to claim 5, in the image signal transmitting method for encoding the prediction residual between the input pixel value and its prediction value, the prediction residual is transferred to the first block. First, it is determined whether or not each of the first blocks is significant based on the activity of the first block.
Of the first block, the first block is subdivided, the second block is generated, and the second block is subjected to the significant judgment. , A quantization step of quantizing the prediction residual included in the block determined to be significant by the step of determining significance for the block and the step of determining significance for the second block, and the quantization output is input. When there is a large amount of buffer remaining in the buffer means that outputs the digitized output at a constant rate, the target value is set based on the remaining buffer so that the target value is increased, and the threshold for significance determination is changed. By predicting the amount of generated information, determining the threshold value for making the transmitted information amount less than or equal to the set target value, and determining the determined threshold value for the first and second blocks. An image signal transmission method characterized by comprising the steps of information amount control using Oite.

The amount of encoded data to be transmitted can be reduced by hierarchically determining the prediction residual, and the amount of encoded data can be efficiently reduced to a target value or less by performing feedback / feedforward buffering. Can be controlled.

[0009]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below with reference to the drawings. A block diagram of an embodiment of the image signal transmission device is shown in FIG. The image data of the current frame is input from the input terminal indicated by 1, and the image data is
It is supplied to the motion detection circuit 2 and the subtractor 3. Further, the prediction residual included in the significant block is quantized by the quantization circuit 4 by the significance determination, and the quantized value is supplied to the inverse quantization circuit 5. In the inverse quantization circuit 5, after the original prediction residual is restored from the quantized value and the threshold value, the prediction residual is supplied to the adder 6.

In the adder 6, the decoded prediction residual and the predicted image data from the prediction circuit 8 are added, that is, the locally decoded output is supplied to the frame memory 7. The frame memory 7 stores the prediction result of the previous frame that has been motion-compensated by local decoding. The image data of the previous frame is supplied from the frame memory 7 to the motion detection circuit 2 and the prediction circuit 8. The motion detection circuit 2 uses the input image data and the image data supplied from the frame memory 7 to detect a motion vector on a block-by-block basis, for example, by block matching between two frames, N frame and (N + 2) frame. Done.

The detected motion vector is supplied from the motion detection circuit 2 to the buffer / framing circuit 17 and the prediction circuit 8. Then, the subtractor 3 subtracts the prediction image data from the prediction circuit 8 from the input image data to generate a prediction residual. The generated prediction residual is supplied to the blocking circuit 10 for hierarchization. In this blocking circuit 10, the supplied prediction residual is 16 lines × 1.
A first block of 6 pixels (hereinafter referred to as 16 × 16 block), a second block of 8 lines × 8 pixels (hereinafter, referred to as
It is divided into blocks of 8 × 8 blocks) and a third block of 4 lines × 4 pixels (hereinafter referred to as 4 × 4 blocks).

The divided 16 × 16 block is supplied from the blocking circuit 10 to the DR table 11, and 8 × 8.
The blocks are fed to the DR table 12 and similarly 4 ×
The four blocks are supplied to the DR table 13. This D
In the R tables 11, 12 and 13, a frequency distribution table of the maximum value (maximum residual) of the absolute value of the prediction residual per frame is created for each layer corresponding to each block. The maximum residual is obtained for each block in each layer.

Further, each DR table 11, 12 and 1
In No. 3, a delay of one frame occurs due to the calculation of the encoded data amount. Then, the frequency distribution table calculated from each of the DR tables 11, 12 and 13 is supplied to the information amount prediction calculation circuit 14. The target bit rate of the encoded data amount is preset in the information amount prediction calculation circuit 14, and the threshold values Th, Th1, T
h2 and Th3 are provided.

In one embodiment of the present invention, feedforward buffering is basically performed. That is,
The information amount predicting / calculating circuit 14 predicts the encoded data amount of one frame, and determines the threshold values Th, Th1, Th2 and Th3 such that the encoded data amount is equal to or less than the target value. With the threshold values Th, Th1, Th2, and Th3 thus determined, significant block determination and quantization processing is performed on the image data of the one frame.

By this buffering, the coded data amount of the frame is made less than or equal to the target value, but the buffer remaining amount used in the next frame of the buffer / framing circuit 17 is the same as that of the buffering process of the previous frame. It depends on the result. For example, when the pattern in the previous frame is at a flat level, the amount of coded data in that frame is small, so the remaining buffer capacity used in the next frame is large and more bits are available for the next frame. Can be given.

Therefore, in the target rate resetting circuit 18,
The target bit rate is set again in consideration of the bit rate and the buffer remaining amount according to the transmission path. The set target bit rate is supplied from the target rate resetting circuit 18 to the information amount prediction calculation circuit 14. In the information amount prediction calculation circuit 14, the preset target bit rate is reset to the target bit rate supplied from the target rate reset circuit 18.

The contents of the threshold ROM 15 are, as will be described later, a first threshold Th for significance determination and second thresholds Th1, Th2, Th for quantized bit allocation.
3 is determined based on measurement, simulation, experience, etc. so as to increase or decrease collectively, and a value is selected so that the buffer / framing circuit 17 does not overflow or underflow.

When the significance judgment circuit 9 compares the threshold Th and the maximum residual and judges the significant block based on the size of the threshold Th, the information amount prediction calculation circuit 14
Then, the threshold Th supplied from the threshold ROM 15
The block with the larger maximum residual is determined to be the significant block.

That is, in the information amount prediction calculation circuit 14,
As will be described later, using the frequency distribution table supplied from each DR table 11, 12 and 13, the coded data amount is predicted according to the threshold value set in the threshold value table,
The threshold number and the mode of hierarchical branching are determined so as to be less than or equal to the target bit rate reset from the remaining buffer capacity.
A significant block is actually determined from this threshold number and the mode of hierarchical branching, and adaptive quantization is performed on the block determined to be a significant block. In addition, the hierarchy number indicating the determined mode of hierarchy branching is supplied from the information amount prediction calculation circuit 14 to the significance determination circuit 9. The threshold number and the layer number at that time are supplied to the buffer and framing circuit 17.

As described above, in the threshold ROM 15, the threshold number is increased by one until the threshold number which is equal to or lower than the target bit rate is determined, and the threshold number is dealt with. Threshold values Th, Th1, Th2, T
h3 is output to the information amount prediction calculation circuit 14.

The significance judging circuit 9 is divided from the information amount predicting / calculating circuit 14 to a hierarchy number for instructing a mode of hierarchy branching, the threshold value ROM 15 to determine the significance threshold value Th, and the blocking circuit 10 to be divided. 16x16 blocks, 8x8 blocks and 4x4 blocks are provided. The significance judgment circuit 9 judges the significance of the block based on the threshold value Th in the form of hierarchy branching based on the hierarchy number, and outputs the significance flag for distinguishing significance / insignificance and the data of the significance block. Is output.

The significance flag from the significance judgment circuit 9 is supplied to the buffer / framing circuit 17, and the data of the significance block is supplied to the quantization circuit 4 as described above. In the quantizer 4, the variable length (semi-fixed length) quantum is supplied to the supplied significant block data based on the threshold values Th1, Th2 and Th3 for quantized bit allocation supplied from the threshold ROM 15. Is done.

The quantizing circuit 4 is an adaptive quantizing circuit having the quantizing characteristics shown in FIG. FIG. 3 shows an example in the case of quantization of 3-bit fixed length including a code. First, when determining a significant block, the maximum residual (dynamic range) is obtained within the block of interest. A value obtained by dividing this maximum residual by 4 is determined as the quantization step width, and the prediction of each pixel is performed. The residual is divided by this quantization step size to be an integer, which is used as a quantization code. Generally, assuming that a prediction residual using ADRC (Dynamic Range Adaptive Coding) with a fixed length of 3 bits is x _i , the number of quantization bits is n, a maximum residual is max, and a quantization code is q _i. ,

It is represented by q _i = [x _i / (max / 2 ^n-1 )]. However, [] means integerization.

Then, the prediction residual of the significant block is quantized in accordance with the determined quantized bit allocation under the control of the quantized bit allocation described later, and the maximum residual, the quantized code, and the significant flag are quantized. Is formatted and transmitted, along with a threshold number and a hierarchy number.

In FIG. 3, since the maximum residual is found on the minus side, the quantization code is (3, 2,
1, 0, -1, -2, -3, -4) has been shown, but when the maximum residual is obtained on the plus side, the 3-bit fixed-length quantization code of this example is: From the plus side, (4, 3, 2, 1, 0, -1, -2, -3). In such quantization, 0 of prediction residual is decoded as 0,
Further, the quantization range is biased toward the polarity where the maximum residual exists, and the quantization error can be reduced.

The quantized value from the quantizing circuit 4 is supplied to the inverse quantizing circuit 5 and the variable length coding circuit 16. In the inverse quantization circuit 5, the quantized value is inversely quantized, that is, the quantized value is decoded, based on the same quantized bit allocation thresholds Th1, Th2, and Th3 supplied to the quantized circuit 4. Is performed, the original prediction residual is restored, and is supplied to the adder 6 as described above. The variable length coding circuit 16 further reduces the amount of data for the supplied quantized value of the significant block by using entropy coding such as Huffman coding, and sends it to the buffer / framing circuit 17 together with the significant flag. Supplied.

The buffer / framing circuit 17 is supplied with a significant flag, a variable-length coded quantized value, a motion vector, and an information amount prediction / calculation circuit 14 with a threshold number and a hierarchical number indicating a mode of hierarchical branching. It The data amount of the quantized value input to the buffer / framing circuit 17 varies depending on the content of the image, but the output of the buffer / framing circuit 17 is output at a constant bit rate according to the capacity of the transmission path.

Here, an example of a hierarchical structure of blocks is shown in FIG.
Shown in The prediction residual is divided into first blocks each having a size of 16 × 16 blocks (first layer: L
1) First, the maximum residual in the target block is detected, and this maximum is compared with a predetermined threshold value. As a result, if the maximum value is smaller than the threshold value, it is determined to be an insignificant block, the significance flag is set to "0", and the prediction residual is not transmitted.

`1 '(significant): (maximum residual)> threshold Th` 0' (insignificant): (maximum residual) ≦ threshold Th

On the other hand, when the maximum residual is larger than the threshold value, it is determined that the block is a significant block, a significant flag is set (it is set to "1"), and the control shifts to the significant determination of the lower hierarchy. . The blocks determined to be significant in the first hierarchy are further subdivided into second blocks having a size of 8 × 8 blocks (second hierarchy: L2), and similar significance determination is performed.
The blocks determined to be significant in the second layer are 4 × 4.
It is divided into third blocks having the size of the block (third layer: L3), and the significance judgment is performed.

Although the hierarchical branching of blocks has been described above as being performed in the order of L1-L2-L3, when the hierarchical branching is performed in the order of L1-L2, the hierarchical branching of L1-L3 is performed. In some cases, hierarchical branching may be performed in order, or hierarchical branching may be performed in L2-L3 order. The information amount prediction calculation circuit 14 determines a mode of hierarchical branching in which the amount of encoded data is the smallest when a certain threshold Th for significance determination is given.

Next, the DR tables 11, 12 and 13
The frequency distribution table of the maximum residuals calculated in 1 will be described with reference to FIG. FIG. 4A shows a frequency distribution table of the maximum residuals in one frame period of the layer L1 configured by 16 × 16 blocks, and the significance determination is performed depending on whether the maximum residuals of the block are larger than the threshold Th. . In this FIG. 4A,
The area s0 on the side smaller than the threshold Th indicates an insignificant block, and the area s1 larger than the threshold Th indicates the number of blocks determined to be significant blocks.

FIG. 4B shows a frequency distribution table of maximum residuals in one frame period of the layer L2 composed of 8 × 8 blocks. This frequency distribution is equal to the block number 4 shown in FIG. 4A.
Double the number of blocks is counted. Further, even if the blocks determined to be significant in the layer L1, there are blocks determined to be significant and blocks determined to be insignificant in the layer L2. In FIG. 4B, the number of blocks determined to be significant is It is represented by the area s2.

Similarly, FIG. 4C shows a frequency distribution table of maximum residuals in one frame period of the layer L3 composed of 4 × 4 blocks. In this frequency distribution, four times as many blocks as the number of blocks shown in FIG. 4B are counted. Further, even in the block of the area s2 that is determined to be significant, there are blocks that are determined to be significant and blocks that are determined to be insignificant in the layer L3. In FIG. 4C, the number of blocks determined to be significant is the area. It is represented by s3.

As described above, these frequency distribution tables are D
It is supplied from the R tables 11, 12 and 13 to the information amount prediction calculation circuit 14, and the information amount prediction calculation circuit 14 obtains the number of significant blocks and the number of significant flag bits for each mode of various hierarchical branches as shown below. To be

In the case of L1 (the number of significant blocks) _L1 = s1 (the number of significant flag bits) _L1 = s0 + s1

In the case of L1-L2-L3 (the number of significant blocks) _L1-L2-L3 = s3 (the number of significant flag bits) _L1-L2-L3 = s0 + s1 + s1 ×
4 + s2 × 4

L1-L3 (Number of significant blocks) _L1-L3 = s3 (Number of significant flag bits) _L1-L3 = s0 + s1 + s1 × 1
6

In the case of L2-L3 (the number of significant blocks) _L2-L3 = s3 (the number of significant flag bits) _L2-L3 = (s0 + s1) × 4 +
s2 x 4

In the case of L1-L2 (the number of significant blocks) _L1-L2 = s3 (the number of significant flag bits) _L1-L2 = s0 + s1 + s1 × 4

The quantizing circuit 4 also performs variable length coding based on the threshold values Th, Th1, Th2 and Th3 for quantized bit allocation. An example of the variable length quantization will be described with reference to FIG. As an example, the number of allocated bits of the adaptive quantizer is set to 2 to 5 bits including 1 bit of the code, and the maximum residual and the quantized bit of the block determined to be the significant block by the threshold value Th for significant determination. Quantized bit allocation is determined as follows by comparing with the allocation thresholds Th1, Th2, and Th3.

Th <(maximum residual) ≦ Th1: 1-bit quantization Th1 <(maximum residual) ≦ Th2: 2-bit quantization Th2 <(maximum residual) ≦ Th3: 3-bit quantization Th3 <(maximum residual Difference): 4-bit quantization

The coded data amount of the coded data by the quantization at this time can be obtained in the same manner as the principle of buffering when using ADRC. The number of quantization bits corresponding to each hierarchical branch (meaning the total number of bits of the quantization code) is determined as follows according to the area of the frequency distribution in the lowest layer in the mode of each hierarchical branch. For example, when only the layer L1 is processed, this layer L1 is the lowest layer, and the layers L1-L2-L
When performing the processing of 3, the layer L3 is the lowest layer.
In the example of FIG. 5, in the frequency distribution table of the third layer, a subscript “3” indicating that the area belongs to the third layer is added to each area. Also for the other layers L1 and L2,
Add 1'and '2' subscripts.

In the case of _L1 (the number of quantization bits) _L1 = (S ₁ 1 × 2 + S ₁ 2 × 3 + S
₁ 3 × 4 + S ₁ 4 × 5) × (16 × 16) + (S ₁ 1+
S ₁ 2 + S ₁ 3 + S ₁ 4) × 8

[0046] When the L1-L2-L3 (number of quantization _{bits) L1-L2-L3 = (} S 3 1 × 2 + S 3 2 ×
3 + S ₃ 3 × 4 + S ₃ 4 × 5) × (4 × 4) + (S ₃ 1
+ S ₃ 2 + S ₃ 3 + S ₃ 4) × 8

In the case of L1-L3 (quantization bit number) _L1-L3 = (S ₃ 1 × 2 + S ₃ 2 × 3)
+ S ₃ 3 × 4 + S ₃ 4 × 5) × (4 × 4) + (S ₃ 1+
S ₃ 2 + S ₃ 3 + S ₃ 4) × 8

In the case of L2-L3 (quantization bit number) _L2-L3 = (S ₃ 1 × 2 + S ₃ 2 × 3)
+ S ₃ 3 × 4 + S ₃ 4 × 5) × (4 × 4) + (S ₃ 1+
S ₃ 2 + S ₃ 3 + S ₃ 4) × 8

In the case of L1-L2 (quantization bit number) _L1-L2 = (S ₂ 1 × 2 + S ₂ 2 × 3)
+ S ₂ 3 × 4 + S ₂ 4 × 5) × (8 × 8) + (S ₂ 1+
S ₂ 2 + S ₂ 3 + S ₂ 4) × 8

When a certain threshold value Th is determined as described above, the number of significant flag bits obtained for each mode of each hierarchical branch is added to the number of quantization bits obtained by the above equation. The encoded data amount of encoded data can be calculated. That is, the coded data amount is calculated by the following equation for each mode of each hierarchical branch. (Amount of encoded data) _* = (Number of significant flag bits) _* +
(Quantization bit number) _* However, * represents a mode of hierarchical branching, and the hierarchical number represents this with, for example, 3 bits.

As described above, the one in which the obtained encoded data amount is the minimum is the most efficient hierarchical branching mode.
Further, the minimum coded data amount is compared with the reset target bit rate, and if the target bit rate is exceeded, the thresholds Th, Th1, Th2, Th3 are changed to set the coded data amount to the target. The same processing is repeated until the bit rate becomes equal to or lower than the bit rate.

Next, R used for the threshold ROM 15
An example of the OM table is shown in FIG. In this ROM table, a threshold for significance determination and a threshold for quantized bit allocation are set, and if the threshold number is small,
The coded data amount is large, and the coded data amount is set to decrease as the threshold number increases, based on measurement, simulation, experience, and the like. is there,
When the target bit rate according to the transmission path is set, the prediction calculation of the encoded data amount is performed as described above from the smaller threshold number according to the processing procedure shown in FIG. The mode of hierarchical branching that produces the smallest amount of encoded data is selected.

After that, the threshold numbers are searched in the order of increasing value, and the same calculation is performed. When the encoded data amount is lower than the target bit rate, the threshold number is determined. By the combination of the threshold values stored in the determined threshold number, significant block determination (including hierarchical branching) and adaptive quantization are actually performed. Finally, in the encoded data to be transmitted, in addition to the above-mentioned significance flag and quantization code, the threshold number and the identification number indicating the mode of hierarchical branching are used as the buffer / framing circuit 17.
Transmitted from.

Here, an example of a procedure for determining the above-mentioned threshold number is shown in FIG. This flow chart is
Starting from step 21, in this step 21, the target bit rate of the encoded data amount to be transmitted is set according to the buffer remaining amount. Then, in step 22, the threshold number is initialized, 0 is set in the threshold number, and control is transferred to step 23. In step 23, it is judged whether or not the threshold number is larger than 15, and if the threshold number is smaller than 15, step 24
If the threshold number is greater than 15, the control proceeds to step 28.

At step 24, the coded data amount is calculated in the manner of each hierarchical branch as described above on the basis of the threshold value by the threshold number. The encoded data amount that is the minimum value is calculated from, and in step 26, it is determined whether or not the encoded data amount is smaller than the target bit rate reset in step 21. If the encoded data amount is larger than the target bit rate, the control moves to step 27, and if the encoded data amount is smaller than the target bit rate, the control moves to step 28.

Then, in step 27, the threshold number is incremented by +1 and control is passed to step 23.
In this way, the threshold number (address) is from 0 to 1.
5, the threshold value for significance judgment and the threshold value for quantized bit allocation are changed each time, and the amount of encoded data (the number of significant flag bits and the number of quantized bits) according to the mode of each hierarchical branch is changed. ) Is calculated. When a threshold number whose encoded data amount calculated by repeating this is smaller than the target bit rate is selected, control is transferred from step 27 to step 28 as described above. And
In step 28, the selected threshold number is determined, and the layer number indicating the mode of the layer branch that caused the encoded data amount at that time is also determined. And
This flowchart ends. Further, in this flowchart, the threshold number and the layer number are determined for each frame.

Here, one embodiment of the decoding side of this image signal transmission apparatus is shown in the block diagram of FIG. The encoded data transmitted from the buffer / framing circuit 17 is supplied to the frame decomposing circuit 32 via the input terminal 31,
The frame decomposing circuit 32 decomposes into a threshold number, encoded data, a motion vector, a significant flag, and a layer number,
Is output. The threshold number is supplied to the threshold ROM 33, and the threshold values Th1, Th2, Th3 for corresponding quantization bit allocation from the ROM table shown in FIG.
Is selected, and the selected threshold values Th1, Th2, Th are selected.
3 is supplied from the threshold ROM 33 to the inverse quantization circuit 35.

The encoded data is supplied to the variable length coding decoder circuit 34, and after the variable length coding is decoded, the inverse quantization circuit 35 supplies the supplied threshold value Th.
Based on 1, Th2 and Th3, the prediction residual is decoded. The decoded prediction residual is 0 in the 0 data interpolation circuit 36 according to the significance flag and the identification number of the hierarchical branch.
It is selected whether to output the data or the supplied prediction residual. The selected output is added to the output from the frame memory 38 in the adder 37.

The motion vector, the significant flag, and the identification number of the hierarchical branch are supplied from the frame disassembling circuit 32 to the address control circuit 39, and the frame memory 38 is controlled based on them. The decoded image of the previous frame is stored in the frame memory 38, the decoded image is output to the adder 37 in accordance with the control signal from the address control circuit 39, and the addition result (that is, the decoded image) is again stored in the frame memory. 38. In addition, the frame memory 38
The decoded image output from and stored in is output via the output terminal 40.

The significant flags are grouped together in a tree structure without compression according to the hierarchical branching mode of the block, or run length Huffman coding is performed on the flags for each hierarchical layer in frame units. The amount of information may be further reduced by carrying out. However, in this case, a frame memory is required.

In this embodiment, the first, second and third frequency distribution tables are generated in parallel. However, after generating the first frequency distribution table, the second frequency distribution table is generated, Furthermore, by subsequently generating the third frequency distribution table, it is possible to appropriately select the mode of hierarchical branching.

Further, in this embodiment, the absolute value of the maximum value of the prediction residual is used as the activity of the block, but the sum of squares of the prediction residuals or the sum of products of the prediction residuals may be used as the activity. It is possible.

Unlike the above-described embodiment, the quantization circuit may perform fixed-length quantization and change the threshold value for significance judgment to control the encoded data amount.

[0064]

According to the present invention, it is possible to improve the compression efficiency by performing the significance judgment in the prediction residual block hierarchically and changing the hierarchical branch of the hierarchy according to the threshold value. In addition, by performing the feedback / feedforward buffering, it becomes possible to control the information amount more efficiently below the target value.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of an image signal transmission device of the present invention.

FIG. 2 is a schematic diagram showing an example of a hierarchical block according to the present invention.

FIG. 3 is a diagram used for explaining an example of linear quantization according to the present invention.

FIG. 4 is a distribution chart showing an example of a frequency distribution table of each layer according to the present invention.

FIG. 5 is a diagram used for explaining an example of nonlinear quantization according to the present invention.

FIG. 6 is an R showing an example of a threshold number according to the present invention.
It is an OM table.

FIG. 7 is a flowchart showing a process of an example of threshold number search according to the present invention.

FIG. 8 is a block diagram showing an embodiment of the decoding side of the image signal transmission device of the present invention.

[Explanation of symbols]

 2 Motion Detection Circuit 3 Subtractor 4 Quantization Circuit 5 Inverse Quantization Circuit 6 Adder 7 Frame Memory 8 Prediction Circuit 9 Significance Judgment Circuit 10 Blocking Circuit 11, 12, 13 DR Table 14 Information Prediction Calculation Circuit 15 Threshold ROM 16 variable length coding circuit 17 buffer and framing circuit

Claims (8)

[Claims] 1. An image signal transmission apparatus for encoding a prediction residual between an input pixel value and its prediction value, wherein the prediction residual is divided into a first block and the first block. Based on the activity of the first block, a significance determining unit for the first block for determining whether or not the first block is significant, and the first block is subdivided to generate a second block, A second block significance determining means for performing the significance determination for every two blocks, and a block determined to be significant by the first block significance determining means and the second block significance determining means. Quantizing means for quantizing the prediction residuals included therein, buffer means for receiving the output of the quantizing means and outputting the output of the quantizing means at a constant rate, and remaining buffer capacity of the buffer means When the number is large, the target value is set based on the remaining buffer so as to increase the target value, and the generated information amount is predicted and transmitted by changing the threshold value for the significance judgment. Is configured to determine the threshold value for setting the amount of information to be equal to or less than the set target value, and to supply the determined threshold value to the significance determining means for the first and second blocks. An image signal transmission device comprising: an information amount control means. 2. The image signal transmission device according to claim 1, wherein the information amount control means forms a frequency distribution table for a predetermined period of the activity of the structures of the first and second blocks, respectively. The number of significant blocks is calculated from the area of a portion larger than a predetermined threshold value for significance determination in the frequency distribution table for each of a plurality of modes of branching from the processing of the first block to the processing of the second block. And predict the generated information amount from the calculated number of significant blocks respectively, compare the predicted minimum value of the generated information amount with a set target value, and predict the minimum value of the generated information amount described above. An image signal transmission device, characterized in that it is configured to determine the above-mentioned threshold value which is equal to or less than a set target value. 3. The image signal transmission apparatus according to claim 1, wherein the quantizing means increases the number of allocated quantized bits when the activity of each of the first and second blocks is large. Is configured so that the quantization step width is controlled, and the information amount control means sets a first threshold value for the significance judgment and a second threshold value for determining the assigned quantization bit number. An image signal transmission device characterized in that the amount of generated information is predicted by changing the amount. 4. The image signal transmission device according to claim 3, wherein the information amount control means forms a frequency distribution table for a predetermined period of the activity of the structure of the first and second blocks, respectively. The number of significant blocks is calculated from the area of a portion of the frequency distribution table that is larger than the first threshold value for each of a plurality of modes of branching from the processing of the first block to the processing of the second block. Then, the generated information amount is predicted from the number of significant blocks calculated for each of the plurality of modes and the number of allocated quantization bits determined by the second threshold value, and the predicted generated information amount is calculated. Is compared with a set target value to determine the first and second threshold values such that the predicted minimum value of the generated information amount is equal to or less than the set target value. Featuring Signal transmission device. 5. An image signal transmission method for encoding a prediction residual between an input pixel value and its prediction value, wherein the prediction residual is divided into a first block and the first block. The step of determining the significance for the first block for determining whether each of the first blocks is significant based on the activity of the first block, and the first block is subdivided to generate a second block. It is determined to be significant by the step of determining the significance for the second block that performs the significance determination for each second block, the step of determining the significance of the first block and the step of determining the significance of the second block. The quantization step of quantizing the prediction residual included in the block, and the quantizer output is input, and the buffer remaining amount of the buffer unit that outputs the quantizer output at a constant rate is If not, the target value is set based on the remaining buffer so as to increase the target value, and the generated information amount is predicted and transmitted by changing the threshold value for the significance judgment. The amount of information to be set is equal to or less than the set target value, the threshold value is determined, and the determined threshold value is used in the significance judgment for the first and second blocks. An image signal transmission method characterized by the above. 6. The image signal transmission method according to claim 1, wherein the step of controlling the amount of information forms a frequency distribution table for a predetermined period of the activity of the structures of the first and second blocks, respectively. For each of a plurality of modes of branching from the processing of the first block to the processing of the second block, the area of a significant block is determined from the area of a portion larger than a predetermined threshold value for significance determination in the frequency distribution table. Calculate the number, predict the generated information amount from the calculated number of significant blocks, respectively, compare the minimum value of the predicted generated information amount with the set target value, and calculate the predicted minimum value of the generated information amount. An image signal transmission method, characterized in that the threshold value is determined so as to be equal to or less than the set target value. 7. The image signal transmission method according to claim 1, wherein in the quantization step, the number of assigned quantization bits increases when the activity of each of the first and second blocks is large. As described above, the quantization step width is controlled, and the information amount control step includes a second threshold value for determining the significance determination first threshold value and the assigned quantization bit number. A method for transmitting an image signal, characterized in that the generated information amount is predicted by varying and. 8. The image signal transmission method according to claim 7, wherein the step of controlling the amount of information forms a frequency distribution table for a predetermined period of the activity of the structure of the first and second blocks, respectively. The number of significant blocks is calculated from the area of a portion of the frequency distribution table larger than the first threshold value for each of a plurality of modes of branching from the processing of the first block to the processing of the second block. And the generated information amount is predicted from the number of significant blocks calculated for each of the plurality of modes and the number of allocated quantization bits determined by the second threshold, and the predicted generated information is calculated. Comparing the minimum value of the amount with a set target value, and determining the first and second threshold values such that the predicted minimum value of the generated information amount is equal to or less than the set target value. Image signal transmission Law.