JPS63296487A - Image signal band compression system, encoding device, and decoding device - Google Patents

Abstract

PURPOSE:To eliminate the need for motion detection on a reception side and to simplify the constitution of the reception side by combining a fixed-density thinning-out process and a variable-density thinning-out process. CONSTITUTION:It is decided on a transmission side whether each area of an image is a motion area or still area, the areas are band-limited correspondingly, and then a thinning-out circuit 2 thins out image elements with fixed density. Further, either of the decision result and the property of the image is used by a thinning-out circuit 3 to thin out the image elements with variable density of the image divided into blocks of predetermined blocks by changing thinning- out and interpolating methods in block units, thereby sending mode information as information on image element signals which are not thinned out and left and the thinning-out and interpolating method. On the reception side, an interpolating circuit 6 performs interpolation corresponding to the thinning-out process of the image elements with the variable density and performs the interpolation with the fixed density corresponding to the motion or still area, thereby restoring the image.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an image signal band compression method and apparatus that can be applied to transmitting or recording image signals.

(Prior art) Conventionally, regarding band compression technology in high-definition video transmission, TA has been reported in IEICE Communications System Study Group 0884-7 "Time axis conversion band compression method for high-definition television signals".
A technique called the T method is known.

The TAT method first extracts the basic pixels representing the basic structure of the image by coarse subsampling, then extracts additional pixels necessary to represent the fineness of the image from among the remaining pixels, and then divides these over a uniform period of time. The signals are rearranged at intervals and transmitted in analog form. This corresponds to dividing five images into small blocks and changing the sampling pattern depending on the fineness of the images within the block. Said 0884-7
In his paper, he proposed a TAT method in which image interpolation is performed using the restored image signal of one frame before in still areas by further performing motion detection, and preprocessing reduces the number of pixels by half. After that, we proposed a method in which the data is further reduced to 1/2 using the TA'I" method, compressing the entire data to 1/4. In this method, in order not to send information about moving parts, both the transmitting and receiving sides Both require motion detection circuitry.

(Problems to be Solved by the Invention) As described above, the conventional TAT method requires motion detection circuits on both the transmitting side and the receiving side, which has the disadvantage that the configuration on the receiving side becomes complicated.

SUMMARY OF THE INVENTION An object of the present invention is to provide an image signal band compression method and apparatus that can eliminate such conventional drawbacks and simplify the configuration of the receiving side by providing a motion detection circuit only on the transmitting side.

(Means for Solving the Problems) In order to solve the above-mentioned problems, the first invention of the present application provides means that, in a band compression method for moving image signals, the transmission side A determination is made as to whether the region is a static region or a static region, and the results of the determination are used to perform band limiting for each of the moving and static regions, and then pixels are thinned out at a fixed density. Using at least one of the following, an image divided into blocks of a predetermined size is thinned out at a variable density by changing the pixel thinning and interpolation method for each block, and the above judgment is performed. The result is not sent to the receiving side, but the remaining pixel signal without being thinned out and mode information, which is information about the thinning and interpolation method, are sent to the receiving side, and on the receiving side,
Using the pixel signal and the mode information, interpolation is performed for pixel thinning at the variable density, and further, based on the mode information, interpolation is performed at a fixed density corresponding to whether each block is a moving area or a static area. It is used to restore images.

Means provided by the second invention of the present application in order to solve the above-mentioned problem is a band compression encoding device for a moving image signal, in which (al) an image signal is input, and each region of the image is divided into a moving region or a still region. (b) a motion/static determination circuit that performs the above determination and outputs the motion/static determination information that is the determination result; (b) inputs an input image signal and the motion/static determination information;
a fixed-density decimation circuit that thins out pixels at a fixed density by performing band limitations corresponding to each of a moving region and a static region; (e) decimating pixels for each block using the output of the fixed-density decimating circuit and the motion/static determination information as input; and (d) a mode determining circuit that determines a mode as an interpolation method and outputs the mode information; (d) a mode determining circuit that uses the output of the fixed density thinning circuit and the mode information as input, and performs variable density processing according to the mode for each block; and the variable density thinning circuit that thins out pixels, and transmits the pixel signals remaining without being thinned out and the mode information to the receiving side.

In order to solve the above-mentioned problem, the means provided by the third invention of the present application determines whether each area of the image is a moving area or a still area for a moving image signal.

Using the results of the determination, band limiting is performed for each of the moving region and the static region, and then pixels are thinned out at a fixed density, and at least one of the results of the determination and the properties of the image are used. Then, an image divided into blocks of a predetermined size is thinned out in blocks, and pixels are thinned out at a variable density by changing the interpolation method. In an image signal band compression decoding device that receives as input the output of an image signal band compression encoder that transmits mode information, which is information regarding the mode, to the receiving side, the transmitting side receives the pixel signal and the mode information as input. (b) A variable density interpolation circuit that performs interpolation corresponding to pixel thinning at the variable density; (b) motion/static determination that uses the mode information as input to determine whether each block is a moving area block or a still area block on a block-by-block basis; a motion and static determination information extraction circuit that extracts information and outputs the extracted motion and static determination information; (C) a motion and static determination information extraction circuit that receives the output of the variable density interpolation circuit and the extracted motion and static determination information as input, and outputs the fixed density on the transmitting side; The fixed-density interpolation circuit performs interpolation corresponding to pixel thinning at , and outputs the restored image signal.

(Operation) In the first invention, the transmitting side first determines whether each area of the image is a moving area or a still area.

Next, using the results of the above-mentioned determination, band limitations are applied to each of the moving region and the static region, and then pixels are thinned out at a fixed density. Furthermore, using at least one of the above judgment result and the image properties, the image divided into blocks of a predetermined size is divided into blocks of a predetermined size, and the pixel thinning and interpolation methods are changed for each block to produce a variable density image. Performs pixel thinning. The result of the determination is not transmitted to the receiving side, but the pixel signals remaining without being thinned out and mode information, which is information regarding the thinning and interpolation method, are transmitted to the receiving side. On the receiving side, interpolation is performed for pixel thinning at the variable density using the pixel signal and mode information, and further, based on the mode information, fixed density pixel thinning is performed for each block, corresponding to whether it is a moving area or a static area. Perform interpolation to restore the image.

The second invention is an apparatus for realizing the transmitting side part of the first invention, and the third invention is an apparatus for realizing the receiving side part of the first invention.

Therefore, the method of the first invention and the method of the second invention. According to the apparatus of the third aspect of the invention, there is no need for a motion detection circuit on the receiving side, and information on whether a moving area block or a stationary area block is used on the receiving side is created from information on the variable density thinning and interpolation method. The amount of information to be transmitted can be smaller than when information about moving area blocks or static area blocks is transmitted to the receiving side.

(Example) Next, an example of the present invention will be described with reference to FIGS. 1 to 11.

FIG. 1 is an overall configuration diagram for clearly showing the configuration of an embodiment of the first invention. An input image signal is inputted from an input terminal 1 to a fixed density thinning circuit 2, a motion/static judgment is made (determination as to whether it is a moving area block or a still area block), band limitation is performed corresponding to each, and pixels are thinned out at a fixed density. The image signal 9 from which the pixels have been thinned out and the motion/static determination information 10 are input to the variable density thinning circuit 3, and pixels are thinned out at variable density using at least one of the motion/static determination result and the image properties. information on what kind of thinning and interpolation method was used, and image information whose pixels have been thinned out.
Type information 11 is transmitted from the output terminal 4 to the receiving side. On the receiving side, the information 11 is inputted from the input terminal 5 to the variable density interpolation circuit 6, and interpolation corresponding to the thinning performed in the variable density thinning circuit 3 is performed according to information on the thinning and interpolation method, and an interpolated image signal 12 is output. . The interpolated image signal and information 13 on the thinning and interpolation method are input to the fixed density interpolation circuit 7, and information on motion and static determination is extracted from the information 13 on the thinning and interpolation method to correspond to the thinning performed in the fixed density thinning circuit 2. interpolation is performed, and the restored image signal 14 is output from the output terminal 8.

FIG. 2 is an overall configuration diagram for clearly showing the configuration of an embodiment of the second invention. An input image signal is input from an input terminal 20 to a motion/static determination circuit 21 and a fixed density thinning circuit 22 . The motion/static determination circuit 21 receives the input image signal, performs motion/static determination (determines whether it is a moving area or a still area), and outputs dynamic/static determination information 27. The fixed-density thinning circuit 22 receives the input image signal and the motion/static determination information 27, performs band limiting corresponding to the motion region and the still region, respectively.
Thin out pixels at a fixed density. Image signal 2 with pixels thinned out
8, motion/static determination information 27, and transmission image signal 29 are input to a mode determining circuit 24, which determines a mode for each block using at least one of the motion/static determination result and the image properties. Output. The variable density thinning circuit 23 receives the image signal 28 and the mode information 30, thins out pixels at a variable density for each block according to the mode information, and outputs a transmission image signal 29. The transmitted image signal 29 and mode information 30 are transmitted from output terminals 25 and 26 to the receiving side, respectively.

FIG. 3 is an overall configuration diagram clearly showing the configuration of an embodiment of the third invention. The transmission image signal 29 and mode information 30, which are the outputs of the encoding device of the second invention, are inputted from input terminals 31 and 32, respectively. The variable density interpolation circuit 33 receives the input image signal 29 and mode information 30, performs interpolation corresponding to variable density thinning on the transmitting side, and outputs an interpolated image signal 38. Mode information 30
is also input to the motion/static determination information extraction circuit 34, which extracts the motion/static determination information 39 from the mode information. The fixed density interpolation circuit 35 inputs the motion/static determination information 39 and the interpolated image signal 38, performs interpolation corresponding to the fixed density thinning on the transmitting side, and outputs a restored image signal from the output terminal 36.

The first, second, and third embodiments of the invention described here are -
For example, as a band compression method, encoding device, and decoding device for television signals, thinning is performed every four fields, first the number of pixels is thinned out to 1/2 of the total by fixed density thinning, and then variable density thinning is performed. We will explain the case where the number of pixels is thinned out to 1/2, the overall number of pixels is thinned out to 1/4 of the total, band compression is performed, and decoding is performed. An example of fixed density thinning is to perform motion/static judgment for each block, limit the band, and thin out pixels. FIG. 6 (b) shows an example of the fixed density thinning method in these embodiments. The same figure (al is the case where pixels are thinned out without applying an offset in odd and even fields, the same figure (b)
This is a case where pixels are thinned out by applying an offset. In variable density thinning, for example, blocks are divided so that the area on the screen is equal to that of blocks in fixed density thinning, and pixels are thinned out by determining a mode for each block. 7th
Figures (al to (dl) are block-by-block thinning and interpolation methods (modes) of variable density thinning in each of these embodiments.
The thinning rate, which is the ratio of the number of thinned pixels to the total number of pixels, is (a) 0, (b) 1/2,
(c), (3 types of dJ3゛/4, (a), (b),
(C) uses 74-rule)" inner sleeve spacing, and (d) uses interframe interpolation using the restored image signal of one frame before. In other words, in this embodiment, in variable density thinning, as an example, FIG. 7(a).

(b), (C1, (d)) 4 Assign the fli type mode to each block, perform thinning corresponding to each block,
It performs band compression. Below, for explanation (al
mode 1, (b) mode 2, tc) mode 3. l
Set dl to mode 4. Further, the size of the block is, for example, 4 lines x 8 pixels in fixed density thinning, and 4 lines/×4 pixels in variable density thinning.

Figure 4 shows the transmission section (Figure 4) of the television signal band compression system which is an embodiment of the first invention. FIG. 6 is a diagram schematically showing the flow of processing of the receiving section (FIG. 5).

As shown in FIG. 4 tar, when an image signal is input (41) on the transmitting side, motion and static determination is first performed for each block (41).
2). For the moving area block, (43) performs band limit corresponding to the moving area (44), and for the stationary area block, performs band limit corresponding to the stationary area (45),
Both are thinned out at a fixed density (46). After fixed-density thinning for all blocks is completed (47), variable-density thinning is performed next.

The method for determining each shade is to first calculate the interpolation error for each mode, and then
5 using the interpolation error for 4 (mode 3, mode 4)
So that the compression rate of the entire image becomes the desired value (1/2)K,
In addition, under the condition that mode 3 and mode 4 are determined according to the motion and static judgment information in fixed density thinning, one entire block is set to thinning rate 0 and thinning rate 3 so that the sum of interpolation errors for the entire image is minimized. /4 divided into two groups. Next, blocks with a thinning rate of 0 and blocks with a thinning rate of 3 are created so that the sum of interpolation errors of the whole image is reduced using the interpolation error while keeping the compression rate of the whole image unchanged.
/4 blocks are changed to a thinning rate of 1/2 (mode 2), and finally the thinning and interpolation methods for all blocks are determined.

The specific flow will be explained using FIG. 4(b). first,
Thinning and interpolation corresponding to each mode are performed (48).

Next, the sum of the absolute values of the differences between the interpolated signal and the original signal within the block, that is, the distortion iD2 for each mode for each block.
．． D3. D4 is calculated (49).

Here, since mode 3 and mode 4 have the same thinning rate of 3/4, selecting either mode has no effect on the compression rate of the entire image. Therefore, according to the result of the motion and static determination performed in Figure 4 fa) 42, mode 3 is selected for the moving area block and mode 4 is selected for the stationary area block for all blocks, and the amount of distortion for the selected mode is set as D5 again. (50, 51, 52, 53
). Next, D is a histogram regarding the amount of distortion D5.
5 Create a histogram (54).

Next, as illustrated in Figure 8 (al), the contents of the D5 histogram are read out and added one after another starting from the one with the largest amount of distortion, and the amount of distortion is calculated when the number of blocks reaches 1/3 or more of the total number of blocks. The value is set as a threshold T1 (55).

All blocks are divided into two groups with a decimation rate of 3/4 and a decimation rate of O, depending on whether the distortion iD5 is smaller than T1, greater than T1, or K.

Next, for a block where one distortion flD5 is smaller than T1, as shown in FIG. 8(b), the distortion amount D5 and the distortion amount D
2 (=D5-D2), and create a histogram for Δ having a value of zero or more. Furthermore, for blocks in which the distortion amount D5 is equal to or greater than T1, as shown in FIG. 8(C), the distortion amount D2
A [D2]1 histogram, which is a histogram for , is created (56).

Furthermore, the threshold T2. Determine T3 (5
7). First, [D2], the blocks that make up the histogram are ordered from the one with the smallest amount of distortion, and D2j (
j=x, 2. a, . . .), and [Δ], the blocks forming the histogram are ordered from the one with the largest amount of distortion, and Δ, (i=1.2, 3, . . .).

Starting from i=l and j==1, the amount of distortion D2J,
D2 for Δh Δ1+1. If there is a relationship of <Δ1+Δ1+1, add 1 to j, add 2 to i, read D2'' again, ΔhΔl+1, repeat this comparison, and the distortion at the time when D2j ≧ Δ1 + Δ51 The value of the amount D2J is set to T3. Let the value of the distortion amount Δ1 be T2. ) and fc) in FIG.
2. Indicates T3.

Next, threshold values Tl, T2. The mode for all blocks is determined using T3 and the result of mode selection by decision 50 (58). That is, for each block, mode 1 is set if D5≧T1 and D2≧T3, and mode 1 is set to D5≧T1 and D2(
If T3, mode 2 is assigned, D5 (TI and Δ), if T2, mode 2 is assigned; and if D5 (TI and Δ≦T2), mode 3 or mode 4 is assigned according to the mode selection result of determination 50.

Finally, pixels of the input image signal are thinned out for each block according to the mode determined in step 58 (59), the pixels are rearranged, and the pixel signal and mode information are output to the receiving side (60). Note that in order to perform interpolation at step 48 corresponding to mode 4 in the next frame, it is necessary to create a restored image signal for the current frame according to the mode determined at step 58.

As shown in FIG. 5, the receiving side inputs the transmitted mode information and pixel signal (61), and performs pixel interpolation at the corresponding variable density based on the received mode information (62).
). Furthermore, motion and static determination information is extracted from the received mode information (63), and for moving area blocks, interpolation is performed at a fixed density for the moving area (65), and for stationary area blocks, interpolation is performed at a fixed density for the stationary area (65). 6
6) Output as a restored image signal (67).

Note that at 48 and 59 in FIG. 4(b), the signal band may be limited in order to prevent aliasing distortion, corresponding to each thinning and interpolation method shown in FIG. The method of band limiting may be adaptively changed between the moving area block and the still area block according to the result of the motion/static determination.

In addition, in FIG. 4 (al), the motion/static determination 42 may be placed in a loop to perform motion/static determination, band limitation, and thinning for one block, and this operation may be performed for all blocks until completion.

In addition, in FIG. 4(b), the distortion calculation 49 is determined by the determination 50.
and D2 and D3 for the dynamic area block.
D2 and D4 may be calculated for the static area block. Furthermore, the D55 histogram generator 54 may be created while calculating D5 (D3 or D4) for each block. .

Furthermore, when determining the threshold value T1 in step 55, the contents of the D5 histogram may be read starting from the one with the smallest amount of distortion. However, in that case, the value of the amount of distortion when the addition result exceeds 2/3 of the total number of blocks is set as T1.

In addition, on the transmitting side, the results of the motion and static determination (42) in fixed density thinning are not used in the mode determination (49 to 58) in variable density thinning, and fixed density thinning and variable density thinning are performed independently. Fixed-density interpolation may be performed by extracting motion and static determination information only from the mode information. However, in this case, 50.51.52
．． When selecting the mode in step 53, for example, the calculated distortion amounts D3 and D for the two types of modes are selected.
4, and select the mode with the smaller amount of distortion.

Alternatively, as a method for determining the mode, the static area block may be set to mode 4 first, and then modes 1 to 3 may be assigned to the moving area block.

The distribution of modes 1 to 3 in this case is shown in Figure 4 (
b) In the mode determination method shown in K, the mode selection in steps 50 to 53 can be carried out without making any fuss, considering D5 = D3. Furthermore, in this case, there is no need to calculate the interpolation error for mode 4, so there is no need to create a restored image signal for the current frame according to the mode determined in step 58.

FIG. 9 is a block diagram schematically showing the basic parts of a television signal band compression encoding apparatus which is an embodiment of the second invention. The second invention relates to the transmitting side device of the first invention. Therefore, as an example, the pixel thinning pattern at fixed density (Figure 6), the type of mode for pixel thinning at variable density (Figure 7), and the mode determination method (Figure 8) are as follows: The same thing as in the embodiment of the invention is used.

The motion/static determination circuit may be implemented using a well-known technique using a one-frame difference. Furthermore, the fixed density thinning circuit may be implemented using well-known techniques of intra-field/in-frame dynamic adaptation.

In the following, the explanation will be centered on the variable density thinning portion.

An image signal is inputted from the input terminal 70, and the motion/static determination circuit 7
1 and a fixed density thinning circuit 72. The motion and static determination circuit 71 performs motion and static determination for each diblock using a known technique and outputs motion and static determination information 95. The fixed density thinning circuit 72 receives the input image signal and the motion/static determination information 95, performs motion/static adaptive band limitation, and outputs an image signal 96 in which the number of pixels has been thinned to 1/2 of the total. The image signal 96 is subjected to mode 2 thinning and interpolation in an interpolation circuit 73 and mode 3 and mode 4 thinning and interpolation in an interpolation circuit 74. The outputs of the interpolation circuits 73 and 74 are input to interpolation error circuits 75 and 76, respectively, and the difference between the interpolation signal and the image signal 96 is calculated.

The difference signal is input to distortion calculation circuits 77 and 78, and the sum of the absolute values of the difference signal within the four blocks, that is, the distortion amounts D2, D3, and D4 for each mode for each block is calculated. Here, the distortion calculation circuit 78 calculates the distortion amounts D3 and D4 for the mode 3 and mode 402 types of modes, but according to the motion/static determination information 95 for each block, mode 3 is used for a moving area block, and mode 3 is used for a static area block. Mode 4 is selected, the amount of distortion D5 for the selected mode is output, and which method of mode 3 or mode 4 is selected is stored, and the mode is determined as information 97 to be referred to at the final mode decision. Output to circuit 88.

The amount of distortion D2 for each thinning rate calculated by the distortion calculation circuits 77 and 78. D5 is each distortion memo IJ 80
, 81 and stored. At the same time, the distortion calculation circuit 7
A histogram of the distortion iD5 is created using the histogram memory 82 by inputting the distortion amount D5 calculated in step 8.

In the threshold value determination circuit 84, as illustrated in FIG. 8(a), the contents of the histogram memory 82 are read out one after another from the side with the largest amount of distortion, and are added up. The value of the amount of distortion at this time is defined as a threshold value T1. T1 is a thinning rate of 0 and a thinning rate of 3/4 for all blocks.
It is used to divide into two groups.

Next is the distortion memo! 780.81, and for blocks where the distortion amount D5 read from the distortion memory 81 is smaller than T1, the distortion amount D5 read from the distortion memory 81 is calculated from the distortion amount D5 read from the distortion memory 81 as shown in FIG. 8(b).
A difference Δ obtained by subtracting the read distortion amount D2 from 0 is calculated in a difference circuit 83, and a histogram of Δ for blocks where Δ is zero or more is created using a histogram memory 86. At the same time, for blocks in which the distortion amount D5 read from the distortion memory 81 is T1 or more, a histogram for the distortion amount D2 read from the distortion memory 80 is created using the histogram memory 85, as shown in FIG. 8 (C1).

Next, in the threshold value determination circuit 87, the threshold value T is determined as follows.
2. Determine T3. Among the contents of the histogram memory 85, the minimum distortion amount among the distortion amounts having frequency data other than O is assumed to be a. Among the contents of the histogram memory 86, the maximum amount of distortion among the amounts of distortion having frequency data that is not 0 is set as b, and the frequency data of the amount of distortion among the contents of the histogram memory 86 is subtracted by 1, and at this point, the histogram memory 86 Let C be the maximum amount of distortion among the amounts of distortion that have frequency data that is not 0 among the contents of . Distortion amount a1 b,
If there is a relationship of a < b + c for h, the frequency data of the distortion amount a from the contents of the histogram memory 85 is subtracted by 1, and the frequency data of the distortion amount C from the contents of the histogram memory 86 is subtracted by 1. do.

The threshold determination circuit 87 repeats the comparison using the above distortion amounts a, b, and c, and sets the value of the distortion amount a at the time when a≧b −1−c as the threshold T3, and the value of the distortion amount S as the threshold T3. Let the threshold value be T2. FIG. 8(b), (C1 shows T2 and T3, respectively.

Next, the mode determining circuit 88 determines the mode for all blocks using the contents of the distortion memories 80 and 81, the values of TI, T2, and T3, and information 97 of the mode determination result in the distortion calculating circuit 78. That is, for each block, if the contents of the distortion memory 81 are d1 and the contents of the strain memory 80 are e, then if d≧T1 and e≧T3, then mode 1 is set; if d≧T1 and e (T3, then mode 2 is set, d (TI and d-e) T2
If d(TI or d-e≦T2), mode 3 or mode 4 is assigned according to the mode determination result in the distortion calculation circuit 78.

After the image signal 96 is output from the fixed density thinning circuit 72 until the mode determining circuit 88 determines the mode for each block, the image signal 96 is subjected to a frame delay 79.
Delay it by one frame.

An image signal 98 that is delayed by one frame in the thinning circuit 89 based on the mode signal 99 that is the output of the mode determining circuit 88
Thin out and rearrange the pixels of .

The thinned-out image signal 100 and the mode signal 99 determined by the mode determination circuit 88 are outputted as output notes 93 and 94, respectively.
sent from to the receiving side.

Further, the image signal 100 thinned out by the thinning circuit 89 and the mode signal 99 determined by the mode determining circuit 88 are inputted to the interpolation circuit 90, where thinning and interpolation corresponding to each mode are performed, and the signals are inputted to the interpolation circuit 74. This is used for mode 4 interpolation as the restored image from one frame before. At the same time, the outputs of decimation and interpolation corresponding to each mode are input to a frame delay 91, delayed by one frame, and then output to an interpolation circuit 9.
When set to 0, it is used for mode 4 interpolation as the restored image of the previous frame.

Note that in the interpolation circuits 73 and 74 and the decimation circuit 89, signal bands may be limited in order to prevent aliasing distortion in accordance with each decimation method shown in FIG.

Further, when determining the threshold value T1 in the threshold value determination circuit 84, the contents of the histogram memory 82 may be read out starting from the one with the smallest amount of distortion. However, in that case, the value of the amount of distortion when the addition result exceeds 2/3 of the total number of blocks is set as TI.

In addition, the motion and static determination information 95 in fixed density thinning is not included in mode determination in variable density thinning,
Fixed-density thinning and variable-density thinning are performed independently, and motion/static determination information is extracted from mode information only on the receiving side to perform fixed-density interpolation. However, in this case, the distortion calculation circuit 78 selects the mode. When performing this, it is necessary to select a mode by, for example, comparing the calculated distortion amounts for two types of modes and selecting the mode with the smaller distortion amount.

Furthermore, the mode selection performed by the distortion calculation circuit 78 may be performed by the interpolation circuit 74. However, in this case, the motion and static determination information 95 is input to the interpolation circuit 74, and the mode selection information 97 is output from the expanded interpolation circuit 74 to the mode determination circuit 88.

Furthermore, the motion and static determination information 95 can also be used as shown in FIG. That is, in FIG. 10, all static region blocks are set to mode 4, and only the distortion memory and histogram memory are used for the motion region blocks to determine the thinning and interpolation method from mode 1 to mode 3. This shows the case where modes are allocated to

Since the motion/static determination information 95 forces the mode determining circuit 106 to set the still area block to mode 4, there is no need to calculate the interpolation error for mode 4. Therefore, the interpolation circuit 101 performs thinning and interpolation for mode 3, and the interpolation error circuit 102 calculates the difference between the interpolation signal and the image signal 96, and the distortion calculation circuit 103 calculates the difference between the interpolation signal and the image signal 96.
The total sum of the absolute values of the difference signals within the block, that is, the distortion amount D3 for mode 3 for each block is calculated and output to the distortion memory 81 and the histogram memory 104. The histogram memory 104 refers to the motion/static determination information 95 and creates a histogram for the distortion amount D3 for only the motion area block. Furthermore, since the restored image signal of one frame before is unnecessary, the local decoder (92 in Fig. 9)
is not necessary. In addition, the threshold determination circuit 105 first calculates the ratio of still area blocks to all blocks using the motion and static determination information 95, and sets all blocks to a thinning rate of 0 and a thinning rate of 0 so that the compression rate of the entire image becomes a desired value. 3/4
After calculating the required number of blocks in mode 3 for dividing into two groups, the contents of the histogram memory 104 are read out one after another from the one with the smallest amount of distortion and added.
The value of the amount of distortion when the number of blocks required for mode 3 is exceeded is defined as a threshold value Tl. At this time, even if the number of blocks in mode 3 is zero, depending on the number of blocks in mode 4, a compression rate higher than the desired compression rate may be obtained. Furthermore, the threshold value T1 can be similarly determined by using the required number of blocks in mode 0 instead of the required number of blocks in mode 3. Further, in the histogram memory 85, a histogram is created regarding the amount of distortion D2 for mode 2 for the moving area block, and in the histogram memory 86, a histogram is created for the amount of distortion D2 for the moving area block.
-D3-D2 Hisdog2m is created. Thereafter, the threshold values T2 and T3 are determined in the threshold value determination circuit 87 in the same manner as in the case of FIG. Mode determination circuit 1
In step 06, mode 4 is assigned to the stationary area block, and mode 1, mode 2, and mode 3 are assigned to the moving area block in the same manner as in FIG. 9 by referring to the motion/static determination information 95. In the case of Fig. 10, the configuration of the receiving side is as follows.
It may be the same as the case in FIG. 9.

FIG. 11 is a block diagram schematically showing the basic parts of a television signal band compression/decoding device which is an embodiment of the third invention. The third invention relates to the receiving side device of the first invention, and as an embodiment of the third invention, an embodiment of the encoding device of the second invention (FIGS. 9 and 10) will be described. ) is received and the image is restored.
This will be explained using Figure 1.

The fixed-density interpolation circuits 118 and 119 may be realized using a well-known technique of intra-field Φ intra-frame dynamic adaptation that corresponds to the decimation circuit on the transmitting side.

The following explanation will focus on the variable density interpolation part in particular.

A reception mode signal 122 and a reception image signal 123 are input from input terminals 110 and 111, respectively.

Interpolation corresponding to each mode shown in FIG. 7 is performed on the image signal 123 by a mode 2 interpolation circuit 112, a mode 3 interpolation circuit 113, and a mode 4 interpolation circuit 114, respectively. A selector 115 selects the signals from each interpolation circuit according to the mode information 122, and outputs a single image signal 124 for restoring an image with variable density thinning. Image signal 12
4 is delayed by one frame by the frame delay 116, and is sent to the mode 4 interpolation circuit 1 as a one frame delay signal 125.
14 is used to perform interpolation corresponding to mode 4.

The mode information 122 is also input to the motion and static determination information extraction circuit 117, and the motion and static determination information 126 is extracted.

As a method for determining this, for example, blocks in mode 4 may be set as static area blocks, and blocks in modes 1 to 3 may be set as moving area blocks.

The image signal 124 that has been subjected to variable density interpolation is then subjected to fixed density interpolation between a moving area and a stationary area using a known technique (1
18, 119), the selector 120 selects the motion/static determination information 1
26 selects the output of either the moving area fixed density interpolation circuit 118 or the static area fixed density interpolation circuit 119 and outputs the restored image signal 127 from the output terminal 121.

According to this embodiment, the signal band of the television signal is 17
4, and the transmission path bandwidth required for transmission is 1/1
4) There is no need to transmit motion/static determination information, and an image signal band compression method with better quality of moving images is possible. Furthermore, if the configuration shown in FIG. 10 is adopted, the restored image signal of one frame before is not required in variable density thinning, and a configuration that does not require a local decoder is also possible when performing interframe interpolation. . Note that the compression rate of the image signal can be varied by controlling the thinning rate in fixed density thinning or by controlling the thinning rate for each block in variable density thinning K or the value of the threshold T1.

(Effects of the Invention) As described above, according to the present invention, since fixed density thinning and variable density thinning are combined, there is no need to transmit motion and static determination information in fixed density thinning.
Furthermore, it is possible to provide an image signal band compression method and an encoding/decoding apparatus that can simplify the configuration of the receiving side by eliminating the need for motion detection on the receiving side.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the basic configuration of the first invention, FIG. 2 is a block diagram showing the basic configuration of the second invention, and FIG. 3 is a block diagram showing the basic configuration of the third invention. FIG. 4 (b) and FIG. 5 are diagrams showing the outline of the processing in an embodiment of the first invention, and FIG. FIG. 7 is an explanatory diagram showing a thinning pattern in fixed density thinning in an embodiment, and FIGS. FIG. 8 (al~(C1) is an explanatory diagram showing a method of determining the thinning and interpolation method in variable density thinning in each embodiment of the present invention, and FIGS. 9 and 10 are one embodiment of the second invention. A block diagram showing an example is shown in Figure 11.
1 is a block diagram showing an embodiment of the invention; FIG. 2.22...Fixed density thinning circuit, 3,23.
...Variable density thinning circuit, 6.33...
Variable density interpolation circuit, 7.35...Fixed density interpolation circuit, 21...Motion/static determination circuit, 24...-
-Mode determination circuit, 34... Motion and static determination information extraction circuit. Agent Patent Attorney Shinsuke HonjoC Figure 5

Claims (4)

[Claims] (1) In the band compression method for video signals, on the transmitting side,
It is determined whether each region of the image is a moving region or a static region, and the results of the determination are used to perform band limiting corresponding to the motion region and the static region, and then pixels are thinned out at a fixed density. Using at least one of the results and the properties of the image, an image divided into blocks of a predetermined size is thinned out at a variable density by changing the pixel thinning and interpolation method for each block. and
The result of the determination is not sent to the receiving side, but the pixel signal remaining without being thinned out and mode information, which is information regarding the thinning and interpolation method, are sent to the receiving side, and the receiving side receives the pixel signal and the mode information. is used to perform interpolation for pixel thinning at the variable density, and further, based on the mode information, interpolation at a fixed density corresponding to whether each block is a moving area or a stationary area is performed to restore the image. Image signal band compression method. (2) In a band compression encoding device for a moving image signal, (a
) A motion/static determination circuit that receives an image signal as input, determines whether each area of the image is a moving area or a static area, and outputs motion/static determination information that is the determination result; (b) A motion/static determination circuit that receives an input image signal and the motion/static determination information As input,
a fixed-density thinning circuit that thins out pixels at a fixed density by performing band limitations corresponding to each in a moving region and a static region; (c) thinning out pixels for each block using the output of the fixed-density thinning circuit and the motion/static determination information as input; and (d) a mode determining circuit that determines a mode as an interpolation method and outputs the mode information; (d) using the output of the fixed density decimation circuit and the mode information as input, and performs variable density processing for each block according to the mode; An image signal band compression encoding device comprising: the variable density thinning circuit that thins out pixels, and transmits the mode information and the pixel signal remaining without being thinned out to a receiving side. (3) The mode determining circuit receives the output of the fixed density thinning circuit, the motion/static determination information, and the output of the variable density thinning circuit as input, and determines a mode as a pixel thinning and interpolation method for each block; The image signal band compression encoding device according to claim (2), characterized in that mode information is output. (4) For the moving image signal, determine whether each area of the image is a moving area or a static area, and use the results of the determination to perform band limiting corresponding to each area of the moving area and static area, and then This is a method of thinning out pixels and interpolating an image divided into blocks of a predetermined size in block units, using at least one of the result of the determination and the properties of the image. Inputs the output of an image signal band compression encoding device that thins out pixels at a variable density by changing the mode, and sends the pixel signal that remains without being thinned out and mode information, which is information about the mode, to the receiving side. In the image signal band compression decoding device, (
a) A variable density interpolation circuit that takes the pixel signal and the mode information as input and performs interpolation corresponding to pixel thinning at the variable density on the transmission side; (b) Takes the mode information as input and performs interpolation for each block in block units; a motion/static determination information extraction circuit that extracts dynamic/static determination information that is information on whether a block is a dynamic area block or a stationary area block, and outputs the extracted dynamic/static determination information; (c) the output of the variable density interpolation circuit and the extraction; and a fixed-density interpolation circuit that receives the motion/static judgment information as input, performs interpolation corresponding to pixel thinning at the fixed density on the transmitting side, and outputs a restored image signal, and outputs the restored image signal. Image signal band compression decoding device.