JPH09238338A - Image signal coding method, image signal coding device, image signal transmission method and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To avoid a compression efficiency from being deteriorated by number of layers even when image data are subject to hierarchical coding and to prevent a low-order layer image from being deteriorated especially. SOLUTION: After layer data D131, D133, D135 are generated by averaging arithmetic operation and picture element data for one picture element among plural lower layer picture elements used for the same averaging operation are excluded from a transmission object to generate one picture element of a higher layer. Then in response to the characteristic of quantization errors when picture element of the highest layer are quantized, the quantization characteristic when the picture element data of the lower layer corresponding spatially to the highest layer picture element are quantized are controlled so that the quantization error has the same polarity as the quantization errors when the highest layer picture element data are quantized. Thus, the hierarchical coding is realized in which the compression efficiency is improved and the deterioration in the image quality is reduced.

Description

Detailed Description of the Invention

[0001]

[Table of Contents] The present invention will be described in the following order. TECHNICAL FIELD OF THE INVENTION Conventional Technology (FIGS. 10 and 11) Problems to be Solved by the Invention Means for Solving the Problems Embodiments of the Invention (1) First Example (FIGS. 1 to 5) (2) ) Second embodiment (FIGS. 6 and 7) (3) Third embodiment (FIGS. 8 and 9) (4) Fourth embodiment (5) Other embodiments Effect of the invention

[0002]

The present invention relates to an image signal coding method, an image signal coding apparatus, an image signal transmission method, and an image signal coding method or apparatus which can be decoded by the image signal decoding apparatus. Regarding a recording medium on which the encoded data is recorded, in particular, predetermined image data is divided into image data of a plurality of layers having different resolutions, and the image data of each layer is encoded to generate encoded data ( That is, it is suitable for application when image data is hierarchically encoded).

[0003]

2. Description of the Related Art Conventionally, in this type of image signal coding apparatus, high resolution input image data is used as first layer image data, and second layer data having a lower resolution than the first layer data, Further, third hierarchical data having a resolution lower than that of the second hierarchical data, ... Is sequentially formed, and each of the plurality of hierarchical data is compression-coded. As a result, hierarchically encoded data for a plurality of layers in which the information amount is sequentially reduced is formed, and the plurality of hierarchically encoded data can be transmitted via a communication path or a recording / reproducing path.

In the image signal decoding apparatus for decoding the plurality of hierarchically encoded data, all of the plurality of hierarchically encoded data can be decoded, and any one of them can be decoded depending on the resolution of the television monitor corresponding to each of them. It is also possible to select and decode a desired one of the hierarchically encoded data of. As a result, if only the desired layer data is decoded from the plurality of layered hierarchical data, the desired image data can be obtained with the minimum required transmission data amount.

Here, as shown in FIG. 10, an image signal coding apparatus 1 which realizes, for example, four-layer coding as the hierarchical coding, has three stages of thinning filters 2 and 3, respectively.
3 and 4 and interpolation filters 5, 6, and 7, and reduced image data D2, D3, and D4 of low resolution in order with respect to the input image data D1 by the thinning filters 2, 3, and 4 of each stage.
And the reduced image data D2, D3, D4 by the interpolation filters 5, 6, 7
Return to 5, D6 and D7 respectively.

Outputs D2 to D4 of the thinning filters 2 to 4
And the outputs D5 to D7 of the respective interpolation filters 5 to 7 are input to the difference circuits 8, 9 and 10, respectively, and the difference data D8, D9 and D10 are generated by the difference circuits. Since the frequency distribution of the difference data D8 to D10 is concentrated near 0, the image coding apparatus 1 can reduce the data amount of the hierarchical data and the signal power. Further, by using run length coding, Huffman coding, or the like in the variable length coding circuit arranged in the subsequent stage, the data amount can be further reduced. Here, the area of each of the difference data D8 to D10 and the reduced image data D4 is 1, 1/4, 1/16, 1/64 with respect to the input image data D1.

The difference data D8 to D10 obtained from the respective difference circuits 8 to 10 and the reduced image data D4 obtained from the thinning filter are provided to the respective encoders 11, 12, 13 ,.
Each of the data is encoded by 14 and compressed. As a result, the first, second, third, and fourth hierarchical data D having different resolutions from the encoders 11, 12, 13, and 14 are obtained.
11, D12, D13 and D14 are sent to the communication path in a predetermined order or recorded on the recording medium via the transmission path.

The first to fourth hierarchical data D11 to D14 thus transmitted are decoded by the image signal decoding device 20 shown in FIG. That is, respectively
The first to fourth hierarchical data D11 to D14 supplied from the communication path or the recording medium via the input terminal are decoded by the decoders 21, 22, 23 and 24, respectively, and as a result, the decoder 24 Outputs the decoded fourth layer data D24.

The output of the decoder 23 is the adder circuit 29.
In, the sum is added to the interpolation data of the fourth hierarchical data D24 obtained from the interpolation filter 26, and thereby the third hierarchical data D23 is restored. Similarly, the decoder 22
Is added to the interpolation data of the third hierarchical data D23 obtained from the interpolation filter 27 in the adder circuit 30 to restore the second hierarchical data D22.
Further, the output of the decoder 21 is added to the interpolation data of the second hierarchical data D22 obtained from the interpolation filter 28 in the adder circuit 31, whereby the first hierarchical data D22 is added.
21 is restored.

[0010]

However, in the image signal coding apparatus 1 which realizes such hierarchical coding, the input image data D1 is divided into a plurality of hierarchical data, and each hierarchical data is coded. Inevitably, the amount of data to be transmitted is increased by the hierarchical component. Therefore, there is a problem in that the compression efficiency is reduced as compared with the high-efficiency coding method that does not use hierarchical coding.

Further, in the image signal decoding apparatus 20, the hierarchical data of the lower layer is restored based on the hierarchical data restored in the upper layer, so that the hierarchical data is sequentially arranged from the upper layer data to the lower layer data. D2
4, D23, D22, D21 are restored. Therefore,
When the encoded upper layer data D14 or D13 is decoded and the decoded data includes an error due to compression encoding, the compression code is included in the restored layer data D21 or D22 of the lower layer where high resolution is originally desired. However, there is a problem that the image quality degradation due to the error is very noticeable in the restored hierarchical data of the lower hierarchy.

The present invention has been made in consideration of the above points, and an image signal encoding method and an image signal encoding method capable of improving compression efficiency when image data is hierarchically encoded and reducing image quality deterioration. An encoding device, an image signal transmission method, and a recording medium in which encoded data encoded by an image signal encoding method or device that can be decoded by an image signal decoding device is recorded is proposed. is there.

[0013]

In order to solve such a problem, in the present invention, a plurality of layers used for the same average value calculation to generate one pixel in the upper layer after each layer data is generated by the average value calculation. The pixel data for one pixel of the lower layer pixels of the above is excluded from the transmission target, and the pixel data of the uppermost layer is spatially divided according to the characteristic of the quantization error when the pixel data of the uppermost layer is quantized. By controlling the quantization characteristic when quantizing the pixel data of the lower layer corresponding to, by controlling the quantization error to be the same as the polarity of the quantization error when quantizing the uppermost layer pixel data, Hierarchical coding with improved compression efficiency and reduced image quality deterioration can be realized.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the drawings.

(1) First Embodiment FIG. 1 shows an image signal coding apparatus 140 according to the first embodiment as a whole. In the first embodiment, a case is shown in which hierarchical data for three layers is formed, and the data for these three layers is compression-coded and transmitted. The image signal encoding device 140 uses the high resolution input image data D.
In the block circuit 141, the block 131 (hereinafter referred to as the first layer image data) is divided into small blocks of 2 lines × 2 pixels, and the block data D13 of the first layer is divided.
2 to form the block data D132 of the first layer
To the decimating circuit 142 and the averaging circuit 143.

The averaging circuit 143 outputs the block data D
The pixel values in each block of 132 are averaged to generate the second layer image data D133 reduced to 1/4 with respect to the input image data D130, and the second layer image data D13 is generated.
3 is sent to the block circuit 144. Similarly to the blocking circuit 141, the blocking circuit 144 divides the second layer image data D133 into small blocks of 2 lines × 2 pixels to form the second layer blocked data D134. The block data D134 of the hierarchical level is sent to the thinning circuit 145 and the averaging circuit 146.

Similar to the averaging circuit 143, the averaging circuit 146 averages the pixel values in each block of the block data D134 to obtain the second layer image data D133.
Reduced to 1/4, that is, the first layer image data D13
Third layer image data D13 reduced to 1/16 with respect to 1
5 is generated. That is, the averaging circuit 143 is configured as shown in FIG.
As shown by the dotted line in (C), 4 in the block of the first layer
Using pixels (for example, X ₁₁ , X ₁₂ , X ₂₁ , X ₂₂ ), the following equation,

[Equation 1] Is performed to generate one pixel (for example, Y ₁₁ ) in the second hierarchy shown in FIG. 2B. Note that the pixels Y ₁₃ , Y
₃₁ ... Are similarly generated by the average of four pixels in the first layer.

Similarly, the averaging circuit 146 is shown in FIG.
4 pixels in the block of the second layer (for example, Y
₁₁ , Y ₁₃ , Y ₃₁ , Y ₃₃ ),

[Equation 2] Is performed to generate one pixel (for example, Z ₁₁ ) in the third layer shown in FIG. 2C. Note that pixels Z ₁₅ and Z
₅₁ , ... Are similarly generated by the 4-pixel average of the second layer.

The thinning circuits 142 and 145 respectively receive the blocked data D132 from the blocking circuit 141 and the blocked data D134 from the blocking circuit 144, and the blocked data D1 composed of four pixels.
One pixel of 32 and D134 is removed, and thinned data D136 and D137 composed of the remaining three pixels excluding the one pixel removed for each block are formed.
Respectively. That is, the thinning circuit 142
Pixels X ₁₁ , X ₁₃ , ... As indicated by the dotted line in FIG. 2C are deleted by thinning, and the thinning circuit 145 generates the pixels in FIG.
Pixels Y ₁₁ , Y ₁₅ , ... As indicated by the dotted line are deleted by thinning.

Therefore, the pixels to be quantized by the first layer quantization circuit 147 and the second layer quantization circuit 148 are the pixels X indicated by the solid lines in FIGS. 2C and 2B.
₁₂ , X ₂₁ , X ₂₂ , ... Or Y ₁₃ , Y ₃₁ , Y ₃₃ ...
Regarding the first layer, the number of pixels to be transmitted can be reduced to 3/4 as compared with the case where all the pixels in the first layer are quantized and transmitted. Also, regarding the second layer, the number of transmitted pixels is reduced to 3/4 compared to the case where all the pixels of the second layer generated by the average calculation are quantized and transmitted.

Further, as a whole, in the image signal coding apparatus 140, as compared with the case of compressing and transmitting only the first layer image data D131, the image data of a plurality of layers with the same number of transmission pixels. It is designed to be able to transmit. As a result, the image signal encoding device 140
Can transmit image data of multiple layers without increasing the amount of transmission information. The thinning circuits 142, 1
The pixels thinned out at 45 can be restored by a decoding side (reception side) described later using a simple arithmetic expression.

The quantizer circuits 147, 148 and 149 are
Each pixel (8 bits) of the thinned-out data D136 and D137 and the third layer image data D135 is, for example, 2
Bits are requantized to compress the amount of information. Further, the quantization circuit 149 of the third layer sends the quantization error information signal S1 indicating the polarity of the error at the time of quantization to the quantization control circuit 200.

The quantizing control circuit 200 sends the quantizing control signals S2 and S3 to the quantizing circuits 148 and 147 of the second and first layers based on the quantizing error information signal S1, and the quantizing circuit 148 and 147 The quantization characteristic is controlled so that the polarity of the quantization error of 147 becomes the same as the polarity of the error of the quantization circuit 149. In this case, the quantization circuit 148
And 147, the lower layer pixel spatially corresponding to the uppermost layer pixel quantized by the quantization circuit 149 is quantized.

This will be specifically described with reference to FIG. FIG. 3 shows that when the input data is 8-bit data per pixel and can take levels from 0 to 255, the quantized value "0" is output to the pixels from 0 to 63 levels. , The quantization value “1” is output to the pixels of the levels of 64-127, the quantization value “2” is output to the pixels of the levels of 128 to 191, and the quantization value is output to the pixels of the levels of 192 to 255. This is a representation of 2-bit quantization that outputs a value "3". This is called the minimum distortion criterion,
This is a method generally used in the past to minimize the quantization error.

In the image signal coding apparatus 140 of the first embodiment, the quantizing circuits 147 and 148 do not use such a minimum distortion criterion, and the polarity of the quantizing error of the quantizing circuit 149, which is the highest layer, is used. Quantization is performed according to. For example, when the pixel data input to the quantizing circuit 149 has a level such as L1 as shown in FIG. 3, the quantizing circuit 149 outputs the quantized value “3” according to the minimum distortion criterion. To do. At this time, the quantization circuit 149 sends the quantization error information signal S1 indicating that the quantization error is + δ1, that is, the positive polarity, to the quantization control circuit 2
Send to 00.

At this time, the quantization circuits 147 and 148
Performs a quantization process such that the quantization error has a positive polarity based on the quantization control signals S2 and S3, not according to the minimum distortion criterion. That is, for example, the quantization circuit 147 or 1
The pixel data input to 48 is L as shown in FIG.
In the case of a level such as 2, the quantized value “2” is output according to the minimum distortion criterion, whereas the quantized circuits 147 and 148 of the first embodiment output from the quantized control circuit 200. The quantized value "3" is output based on the quantized error information signal S1 (in this case, S1 represents the positive polarity). Further, for example, when the input pixel data has a level such as L3, based on the quantization error information signal S1 from the quantization control circuit 200,
A quantized value "2" is output. This is the same as when the minimum distortion criterion is followed.

Requantized data D138, D139, D14 obtained by the quantization circuits 147, 148, 149.
0 is a variable length coding circuit (VLC) 150, 1 respectively
51, 152. Variable length coding circuit 150-
Each of 152 includes requantized data D138 to D14.
A shorter Huffman code is assigned to 0 for a quantized code having a higher occurrence frequency, and each requantized data D138 to D138
First layer coded data D141, second layer coded data D142, and third layer coded data D143, each of which represents 140 with a code amount as small as possible, are formed. Hierarchical coding D1
42 and the third layer coding D143 are sent to the transmission format conversion circuit 153.

The transmission format conversion circuit 153 has a first
Layered coded data D141, second layered coded data D1
42 and the third layer encoded data D143 are arranged in a predetermined order, or an identification code for identifying the layer of each layer encoded data is added to the transmission image data D
144 is formed, and this transmission image data D144 is output. This output transmission image data D144 is then supplied to the receiving side via the communication path 154 or recorded on the recording medium 155 such as a disk, tape or semiconductor memory via the recording transmission path.

The image signal decoding device 160 for decoding the transmission image data D144 thus formed can be constructed, for example, as shown in FIG. The image signal decoding apparatus 160 inputs the transmission image data D144, which is supplied through the communication path 154 or reproduced from the recording medium 155 through the reproduction transmission path, to the data shunt circuit 161. The data shunting circuit 161 has a switching circuit (not shown) and refers to the identification code of each layer included in the transmission image data D144 to convert the transmission image data D144 into the first layer encoded data D150 and the second layer encoded data D151. And third layer encoded data D152
To the first layer encoded data D150, the second layer
Layered coded data D151 and third layered coded data D
152 are variable length decoding circuits (IVLC) 16
2, 163, 164.

Variable length decoding circuits 162, 163, 164
Performs the reverse processing of the variable length coding circuits 150, 151 and 152 shown in FIG. 1 described above, respectively, and the first layer coded data D15 represented by the Huffman code.
0, the second layer coded data D151, and the third layer coded data D152 are converted into requantized data D153, D154, and D155 represented by the requantized code. Then, these requantized data D153, D15
4, D155 are inverse quantization circuits 165, 166,
167.

The dequantization circuit 165 dequantizes the requantized data D153, which is set to 2 bits per pixel by the quantization circuit 147 shown in FIG. One-layer thinned-out decoded data D156 is generated. Then, the first layer thinned-out decoded data D1
56 is sent to the synthesis circuit 168 and the pixel generation circuit 169. The dequantization circuit 166 dequantizes the requantized data D154, which is set to 2 bits per pixel by the quantization circuit 148 shown in FIG. Decoded data D157 is generated. Then, the second layer thinned-out decoded data D157 is sent to the synthesizing circuit 170 and the pixel generating circuit 171.

The dequantization circuit 167 dequantizes the requantized data D155, which is 2 bits per pixel by the quantization circuit 149 shown in FIG. To the third layer decoded image data D15
Output as 8. Then, the third-layer decoded image data D158 is sent to a low-resolution television monitor with a small number of display pixels and also to the pixel generation circuit 171, through the output terminal, for example.

The pixel generation circuit 171 uses the third-layer decoded image data D158 and the second-layer thinned-out decoded data D157, and the image signal coding apparatus 1 shown in FIG.
The second layer pixels (that is, the pixels indicated by the dotted line in FIG. 2B) thinned out by the thinning circuit 40 of 40 are restored. For example, the pixel Y ₁₁ of the second layer deleted by thinning is

(Equation 3) Restoration is performed by performing an operation such as. Similarly, all the thinned pixels are restored by using the upper layer pixels generated by averaging and the pixels that are used for generating the upper layer pixels and are not thinned.

The synthesizing circuit 170 inserts the decompressed pixel data D159 of the second layer thus generated into a predetermined position in the second layer thinned-out decoded data D157 and synthesizes the second layer decompressed image. The data D160 is formed. Then, the second layer decoded image data D160 is sent to a pixel monitor circuit 169 as well as sent to a television monitor having a medium number of display pixels through an output terminal, for example.

The pixel generation circuit 169 uses the second-layer decoded image data D160 and the first-layer thinned-out decoded data D156, and the image signal coding apparatus 1 shown in FIG.
The first layer pixels (that is, the pixels indicated by the dotted line in FIG. 2C) thinned by the thinning circuit 142 of 40 are restored. For example, the first layer pixel X ₁₁ deleted by thinning is

(Equation 4) Restoration is performed by performing an operation such as. Similarly, all the thinned pixels are restored by using the upper layer pixels generated by averaging and the pixels that are used for generating the upper layer pixels and are not thinned.

The synthesizing circuit 168 inserts the reconstructed pixel data D161 of the first layer thus generated into a predetermined position in the first layer thinned-out decoded data D156 and synthesizes the first layer decoded image. The data D162 is formed. Then, the first layer decoded image data D162 is sent to a high-vision television monitor having a large number of display pixels, for example, through an output terminal.

The image signal encoding apparatus 1 having the above-mentioned configuration
The image data D1 of the plurality of layers 40 is generated by generating the upper layer pixels by the average value of the plurality of pixels of the lower layer.
31, D133, and D135 are generated.

In addition to this, the image signal coding device 140
Is excluded from the transmission pixels because one pixel of the pixels used for the same averaging operation can be restored by a simple arithmetic operation on the decoding side in the hierarchical image data excluding the highest hierarchy, that is, the third hierarchy. As a result, the image signal encoding device 14
In 0, it is possible to realize the hierarchical encoding process without increasing the number of transmission pixels due to the hierarchical structure.

By the way, in the image signal coding apparatus 140, the requantization circuits 147 to 149 requantize each layer image data to compress the data amount in each layer. As a result, each piece of the most quantized data D138, D13
9, D140 inevitably becomes a value including a quantization error at the time of requantization. As the quantization error increases, the decoded image data D158, D160, and D162 have a larger error than the true value, and the image quality deteriorates.

Therefore, the influence of the quantization error in each layer in the image signal coding device 140 and the image signal decoding device 160 will be considered. Here, each pixel Z ₁₁ , Y ₁₁ , ...
The decoded value of ... is Z ₁₁ ′, Y ₁₁ ′, ... And the true value is Z ₁₁ ,
Y ₁₁ , ..., And the quantization error is E (Z ₁₁ ), E (Y ₁₁ ).
.., the decoded value Z ₁₁ ′ of the pixel Z ₁₁ of the third layer obtained by the inverse quantization circuit 167 is expressed by the following equation,

(Equation 5) Becomes In addition, the decoded values Y ₁₃ ′ of the pixels Y ₁₃ , Y ₃₁ , and Y ₃₃ of the second layer obtained by the inverse quantization circuit 166,
Y ₃₁ ′ and Y ₃₃ ′ are the following equations,

(Equation 6) Becomes

However, since the second layer pixel Y ₁₁ restored by the pixel generation circuit 171 is generated based on the equation (3), its decoded value Y ₁₁ ′ is given by the following equation:

(Equation 7) Therefore, the quantization error with respect to Z ₁₁ is affected by 4 times.

Also, the decoded values X ₁₂ ′, X ₂₁ ′, X ₂₂ ′ of the first layer pixels X ₁₂ , X ₂₁ , X ₂₂ , X ₁₄ , X ₂₃ , X ₂₄ obtained by the inverse quantization circuit 165. , X ₁₄ ′, X ₂₃ ′,
X ₂₄ ′ is the following equation,

(Equation 8) However, the first is restored by the pixel generation circuit 169.
Since the hierarchical pixels X ₁₃ , X ₃₁ , and X ₃₃ are generated based on the equation (4), their decoded values X ₁₃ ′, X ₃₁ ′, and X ₃₃ ′ are as follows:

[Equation 9] And each decoded value X ₁₃ ′, X ₃₁ ′, X ₃₃ ′,
The quantization error for the spatially corresponding second layer pixels Y ₁₃ , Y ₃₁ , Y ₃₃ , that is, the corresponding upper layer pixels, is affected by a factor of four.

Further, in the first layer pixel X ₁₁ which is restored by the pixel generating circuit 169 via the pixel generating circuit 171, the decoded value X ₁₁ ′ is as follows:

(Equation 10) , The decoded value X ₁₁ ′ is added to the second layer pixels Y ₁₃ , Y 2.
_In addition to the quantization error of ₃₁ and Y ₃₃ being affected by 4 times, the quantization error of the third layer pixel Z ₁₁ being affected by 16 times.

FIG. 5 shows that each decoded pixel value Z ₁₁ ′,
Quantization error E for Y ₁₁ ′, Y ₁₃ ′, ...
(Z ₁₁ ), E (Y ₁₃ ), E (Y ₃₁ ), ... As you can see in Figure 5,
The quantization error in the upper layer has a great influence on the decoded value in the lower layer.

In consideration of this, in the first embodiment of the present invention, the pixel Z ₁₁ of the highest hierarchy is spatially divided into pixels of the highest hierarchy Z ₁₁ based on the polarity of the quantization error when the pixel Z ₁₁ of the highest hierarchy is quantized. Corresponding lower hierarchical pixels Y ₁₃ , Y ₃₁ , Y ₃₃ , X ₁₂ ,
The quantization of X ₂₁ and X ₂₂ is controlled so that the polarity of the quantization error becomes the same as the polarity of the quantization error when the pixel Z ₁₁ of the highest hierarchy is quantized. That is, for example, as is clear from equation (7), when decoding the second hierarchical pixel Y ₁₁ by the reverse process of the average calculation, quantization error E (Z ₁₁₎ of the Z ₁₁ is an upper hierarchy pixel and the Quantization errors E (Y ₁₃ ), E (Y ₃₁ ), and E (Y ₃₃ ) of Y ₁₃ , Y ₃₁ , and Y ₃₃ which are hierarchical (second hierarchical) pixels (pixels existing in the same block) are included. It will be. However, the upper layer pixel Z
₁₁ is a same hierarchical pixel as the quantization error E (Z ₁₁₎ of Y _13, Y
_31, the quantization error E (Y ₁₃₎ of _{Y 33, E (Y 31)} , E (Y
The relationship with ₃₃ ) is a canceling relationship, as is clear from equation (7).

[0046] Yotsute, Y _13, Y _31, the quantization error E Y ₃₃ (Y ₁₃₎ which is the same hierarchical pixel as the quantization error E (Z ₁₁₎ of the Z ₁₁ is an upper hierarchy pixel, E (Y ₃₁ ) And E (Y ₃₃ ) have the same polarity, it is possible to reduce the quantization error during decoding. Therefore, according to the first embodiment of the present invention, as is clear from the expression (7), the lower layer is quantized so that it has the same polarity as the uppermost layer at the time of encoding, Sometimes the quantization errors cancel each other out, and it is possible to reduce image quality deterioration in the lower hierarchy.

Similarly, as is apparent from the equation (10), when the first layer pixel X ₁₁ is decoded by the inverse process of the averaging operation, the highest layer (third layer) pixel Z ₁₁ Quantization error E (Z ₁₁ ) and Y, which is the upper layer (second layer)
Quantization error E (Y ₁₃ ), E (Y ₁₃ ), of ₁₃ , Y ₃₁ , Y ₃₃ ,
Quantization errors E (X ₁₂ ), E (X ₂₁ ), E of E (Y ₃₃ ) and X ₁₂ , X ₂₁ , and X ₂₂ which are pixels of the same layer (first layer) (pixels existing in the same block) (X ₂₂ ) will be included. However, the quantization error E (Z ₁₁ ) of Z ₁₁ which is the highest layer pixel and Y ₁₃ , Y ₃₁ , Y which is the upper layer
₃₃ of the quantization error _{E (Y 13), E (} Y 13), E (Y 33) and X _12, X _21, quantization error E X ₂₂ is a same-layer pixel
The relationship between (X ₁₂ ), E (X ₂₁ ), and E (X ₂₂ ) is (1
As is clear from the equation (0), there is a canceling relationship.

Therefore, the quantization error E (Z ₁₁ ) of Z ₁₁ which is the highest layer pixel and Y ₁₃ , Y ₃₁ , Y ₃₃ which is the upper layer.
The quantization error _{E (Y 13), E (} Y 13), E (Y 33) and X _12, X _21, the X ₂₂ quantization error E is equal hierarchical pixel (X
₁₂ ), E (X ₂₁ ), and E (X ₂₂ ) are encoded so that they all have the same polarity, it is possible to reduce the quantization error during decoding. Therefore, according to the first embodiment of the present invention, as is clear from the equation (10), the lower layer is quantized so that it has the same polarity as the uppermost layer at the time of encoding, Sometimes the quantization errors cancel each other out, and it is possible to reduce image quality deterioration in the lower hierarchy.

As described above, in the image signal coding apparatus 140, the decoding side can perform the quantization process in consideration of the quantization error propagating from the upper layer to the lower layer, and the image quality deterioration based on the quantization error. Can be reduced.

According to the configuration of the first embodiment described above, the pixel data Y ₁₁ for one pixel of the plurality of lower layer pixels used for the same average value calculation to generate one pixel in the upper layer,
Y ₁₅ , ..., X ₁₁ , X ₁₃ , ... Are not transmitted, and the lower layer corresponding to the uppermost layer pixel is determined according to the polarity of the quantization error when the uppermost layer pixel is quantized. Improves compression efficiency by controlling the quantization characteristics when quantizing hierarchical pixels so that the quantization error has the same polarity as the quantization error when quantizing the highest hierarchical pixel. It is possible to reduce deterioration of image quality.

(2) Second Embodiment FIG. 6 shows, as a whole, an image signal coding apparatus 1 according to the second embodiment.
80 is shown. Compared with the first embodiment, it has the same configuration as the image signal encoding apparatus 140 of FIG. 1 except that the residual (difference) compression encoding of inter-layer data is performed except for the highest layer. Therefore, the parts corresponding to those in FIG.

That is, the image signal coding apparatus 140 of the first embodiment sets each pixel to PCM (Pulse Code Modulation).
While the image signal encoding device 180 of the second embodiment transmits each pixel by D
Transmission is performed in the form of PCM (Differential Pulse Code Modulation). Therefore, the image signal encoding device 180 can further reduce the amount of transmission information.

More specifically, the image signal encoding device 180 supplies the third layer image data D135 and the second layer blocked data D134 to the difference circuit 181.
Then, the third level image data D is calculated by the difference circuit 181.
The difference between pixels spatially corresponding to each other between 135 and the second layer block data D134 is calculated to form the second layer difference data D170.
70 is sent to the thinning circuit 182. At this time, the difference circuit 181 determines the difference value Δ of the second layer pixels Y ₁₃ , Y ₃₁ , and Y _33.
Y ₁₃ , ΔY ₃₁ , and ΔY ₃₃ are calculated by using the upper layer pixel Z ₁₁ corresponding to these pixels,

[Equation 11] To ask.

The image signal encoding device 180 similarly supplies the second layer blocked data D134 and the first layer blocked data D132 to the difference circuit 183.
The difference circuit 183 is used to output the second layer block data D1.
34 and the first layer block data D132, the difference between pixels spatially corresponding to each other is calculated to form the first layer difference data D171.
1 is sent to the thinning circuit 184. At this time, the difference circuit 183 determines the difference value ΔX of the first layer pixels X ₁₂ , X ₂₁ , and X _22.
12 , ΔX ₂₁ , ΔX ₂₂ are calculated by using the upper layer pixel Y ₁₁ corresponding to these pixels,

(Equation 12) To ask.

The thinning circuits 184 and 182 are similar to the thinning circuits 142 and 145 shown in FIG. 1 described above, and the blocked data D132 from the blocking circuit 141 and the blocked data D134 from the blocking circuit 144.
Respectively. Then, one pixel is thinned out from each of the first layer and second layer difference data D171 and D170 composed of four pixels for each block corresponding to the block data D132 and D134 composed of four pixels. Hierarchical difference subtraction data D1 composed of the remaining three pixels excluding the one pixel removed for each block
72, D173 is formed, and this layer difference subtraction data D
172 and D173 are sent to the quantization circuits 185 and 186.

The third layer quantizing circuit 187 receives the third layer image data from the averaging circuit 146, and in the same manner as in the first embodiment, each pixel follows the distortion minimum criterion, for example, 2 pixels.
Quantize to bits. Then, the quantization circuit 187 sends a quantization error information signal S11 indicating whether the quantization error has a positive polarity or a negative polarity to the quantization control circuit 210.

Here, the quantization circuit 185 of the first layer and the quantization circuit 186 of the second layer are the same as in the first embodiment.
The quantization characteristic is controlled based on the quantization control signals S12 and S13 output from the quantization control circuit 210 according to the polarity of the quantization error of the quantization circuit 187 in the third layer (top layer). That is, when the polarity of the quantization error by the quantization circuit 187 of the upper layer pixel is positive, the quantization circuit 185 for quantizing the lower layer pixel spatially corresponding to this upper layer pixel, The quantization characteristic of 186 is controlled so that the polarity of the quantization error is positive. When the quantization error of the quantization circuit 187 of the upper layer pixel is negative, the quantization circuit 185 for quantizing the lower layer pixel spatially corresponding to the upper layer pixel, The quantization characteristic of 186 is controlled so that the polarity of the quantization error is negative.

Requantized data D174, D175, D176 obtained by the quantization circuits 185, 186, 187.
Is a variable length coding circuit (VLC) 150, 151, 15
2 is variable-length coded and is output as first layer, second layer, and third layer encoded data D177, D178, and D179. Then, these first layer, second layer and third layer encoded data D177, D178, D17
9 is input to the transmission format conversion circuit 153,
The transmission format conversion circuit 153 transmits the transmission image data D
Form 180 and output it. The output transmission image data D180 is then supplied to the receiving side via the communication path 188, or the recording medium 189 such as a disk, a tape or a semiconductor memory via the recording transmission path.
Will be recorded.

The image signal decoding device 190 for decoding the transmission image data D144 thus formed can be constructed, for example, as shown in FIG. In the image signal decoding device 190, the same components as those of the image signal decoding device 160 shown in FIG.
The image signal decoding device 190 is supplied via the communication path 188 or is supplied to the recording medium 18 via the reproduction transmission path.
The transmission image data D180 reproduced from No. 9 is input to the data shunt circuit 161. The data diversion circuit 161 diverts the transmission image data D180 into first layer, second layer and third layer encoded data D181, D182 and D183, and these first layer, second layer and third layer encoded data. D181, D182 and D183 are supplied to variable length decoding circuits (IVLC) 162, 163 and 164, respectively.

Variable length decoding circuit (IVLC) 162, 1
62 and 164 respectively represent the first layer, the second layer and the third layer encoded data D181, D182 and D183.
Of the requantized data D184, D
185 and D186, and these requantized data D
184, D185, and D186 are inverse quantization circuits 191, 1
To 92 and 193 respectively. The variable length decoding circuits (IVLC) 162, 163 and 164 are shown in FIG.
Variable length coding circuits (VLC) 150 and 151 shown in FIG.
Inverse processing corresponding to each of Nos. And 152 is executed.

Inverse quantization circuits 191, 192 and 193
Is 8 bits of requantized data D184, D185 and D186 which are requantized to 2 bits per pixel.
Inversely quantized into bit data, and subtracted data D187 and D18 of the first layer and the second layer
8 and the third layer decoded image data D189 are generated respectively. The inverse quantization circuits 191, 192 and 193
Are quantization circuits 185, 186 and 18 shown in FIG.
Reverse processing corresponding to 7 is executed.

The third layer decoded image data D189 is directly output to a low resolution television monitor or the like, and is also output to the adder circuit 194 and the pixel generation circuit 195. The adder circuit 194 has the following equation:

(Equation 13) And the second layer thinned-out decoded data D
Calculate 190.

The pixel generation circuit 195 uses the second layer difference / batch data D188 and the third layer decoded image data D189 to extract the second layer pixel Y ₁₁ thinned by the thinning circuit 182 shown in FIG. , The following equation,

[Equation 14] Ask by In the equation (14), when compared with the above-mentioned equation (3), the multiplication coefficient of the third layer pixel Z ₁₁ is “4” in the equation (3), whereas it is “1” in this equation. You can see that it is running. This means that in equation (14),
It means that the influence of the pixel value of the third layer on the pixel value of the second layer is smaller than that of the expression (3).

The synthesizing circuit 196 synthesizes the second layer thinned-out decoded data D190 and the second layer restored pixel data D191 to form the second layer decoded image data D192, and outputs the second layer decoded image data D192. Output to a television monitor or the like via the adder circuit 197.
And to the pixel generation circuit 198. Adder circuit 197
Is:

(Equation 15) The addition operation as shown in FIG.
193 is calculated.

The pixel generation circuit 198 uses the first layer difference / batch data D187 and the second layer decoded image data D192 to thin the first layer pixel X ₁₃ , which is thinned by the thinning circuit 184 shown in FIG. X ₃₁ and X ₃₃ are given by

(Equation 16) Ask by In addition, the pixel generation circuit 198 converts the thinned-out first layer pixel X ₁₁ into the following equation:

[Equation 17] Ask by

The synthesizing circuit 199 synthesizes the first layer thinned-out decoded data D193 and the first layer restored pixel data D194 to form the first layer decoded image data D195.
Then, the first layer decoded image data D195 is output to, for example, a high-resolution television monitor or the like via the output terminal.

Next, as in the first embodiment, the influence of the quantization error in each layer in the image signal coding device 180 and the image signal decoding device 190 of the second embodiment will be considered. The decoded value Z ₁₁ ′ of the pixel Z ₁₁ in the third layer obtained by the inverse quantization circuit 193 is the same as that in the expression (5). The decoded values Y ₁₃ ′, Y ₃₁ ′, and Y ₃₃ ′ of the pixels Y ₁₃ , Y ₃₁ , and Y ₃₃ of the second layer obtained by the adder circuit 194 are
The following formula,

(Equation 18) It becomes the value shown by.

Since the second layer pixel Y ₁₁ restored by the pixel generation circuit 195 is generated based on the equation (14), its decoded value Y ₁₁ ′ is

[Equation 19] It becomes the value shown by. As described above, the image signal encoding device 180 of the second embodiment transmits the difference data between the layers as the transmission data, excluding the highest layer data, and (1
Comparing the equation (9) with the above-mentioned equation (7), the multiplication coefficient of the quantization error E (Z ₁₁ ) of the third layer pixel Z ₁₁ is “4” in the equation (7). You can see that it is 2 ”.

This means that according to the coding and decoding of the second embodiment, the influence of the quantization error of the third layer pixel on the decoding of the second layer pixel can be reduced to about half. To do. That is, in the first embodiment, (1
As is clear from the equation (9), since the difference between the layers is calculated, the quantization error of the pixel Z ₁₁ of the third layer is the decoded pixel values Y ₁₁ ′, Y ₁₃ ′, Y of the second layer. It is reflected in all of ₃₁ 'and Y ₃₃ '. Therefore, the influence of the quantization error of the third layer pixel on the decoding of the second layer pixel can be reduced.

The first obtained by the adder circuit 197
The decoded values X ₁₂ ′, X ₂₁ ′, X ₂₂ ′, X ₃₂ ′, X ₄₁ ′, X ₄₂ ′ of the hierarchical pixels X ₁₂ , X ₂₁ , X ₂₂ , X ₃₂ , X ₄₁ , X ₄₂ are:
The following formula,

(Equation 20) And the decoded value X ₃₁ ′ of the first layer pixel X ₃₁ restored by the pixel generation circuit 198 is

(Equation 21) The value is indicated by.

Further, the decoded value X ₁₁ ′ of the first layer pixel X ₁₁ restored by the pixel generation circuit 198 via the pixel generation circuit 195 is given by the following equation:

(Equation 22) The quantization error for Z ₁₁ has an effect of 4 times, and the quantization error for Y ₁₃ , Y ₃₁ , Y ₃₃ has an effect of 2 times. However, this (22)
As is clear by comparing the equation and the equation (10), Z ₁₁ , Y
_The influence of the quantization error of ₁₃ , Y ₃₁ and Y _{33 on} the decoded value X ₁₁ ′ is significantly reduced as compared with the case of the first embodiment. That is, in the first embodiment, for example, as is clear from the equation (10), the quantization error of the pixel Z _{11 in} the third layer and the second
The quantization error of the pixels Y ₁₃ , Y ₃₁ , Y _{33 in} the hierarchy is reflected only in the decoded pixel value X ₁₁ ′ in the first hierarchy, and the decoded pixels X ₁₂ ′, X ₂₁ ′, X _{22 in} the first hierarchy are reflected. Not reflected in ′.

However, in the second embodiment, (2
As is clear from the equation (2), since the difference between the layers is calculated, the quantization error of the pixel Z ₁₁ of the third layer and the quantization error of the pixels Y ₁₃ , Y ₃₁ , and Y ₃₃ of the second layer. Will be reflected in the decoded pixels X ₁₁ ′, X ₁₂ ′ and X ₂₁ ′ of the first layer. Therefore, the influence of the quantization error of the third layer pixel and the quantization error of the second layer pixel on the decoding of the first layer pixel can be reduced.

Therefore, according to the configuration of the image signal coding apparatus of the second embodiment, each layer data of a plurality of layer image data obtained by the average value calculation from the input image data D131 and the adjacent upper layer. In the image signal encoding device 180 that generates hierarchical difference data with respect to the data, and quantizes the highest hierarchy data D135 and the plurality of hierarchy difference data D172 and D173, respectively, to generate a plurality of hierarchy coded data. Difference pixel data ΔY ₁₁ , ΔY ₁₅ , ... Which can be restored by arithmetic operation using the pixels of the adjacent lower layer and the pixels of its own layer among the pixels of each layer except
Not to transmit ΔX ₁₁ , ΔX ₁₃ , ...
According to the polarity of the quantization error when the uppermost layer pixel is quantized, the quantization characteristic when quantizing the lower layer pixel corresponding to the uppermost layer pixel is By controlling so as to have the same polarity as the quantization error at the time of quantization, the compression efficiency can be further improved while suppressing the deterioration of image quality.

(3) Third Embodiment FIG. 8 shows the overall image signal coding apparatus 4 of the third embodiment.
00 is shown. The image signal encoding device 400 is
Compared to the second embodiment, adaptive prediction circuits 401 and 402
The configuration is the same as that of the image signal encoding device 180 of the second embodiment except that is provided. Therefore, the parts corresponding to those in FIG. 6 are designated by the same reference numerals.

The adaptive prediction circuit 401 performs a predetermined prediction process based on the third layer image data D135, and the second layer prediction data D4 corresponding to the second layer image data D134.
00 is generated. Then, the second layer prediction data is sent to the difference circuit 181. Similarly, the adaptive prediction circuit 402
Performs a predetermined prediction process based on the second layer image data D134 and generates first layer prediction data D401 corresponding to the first layer image data D132. Then, the first layer prediction data is sent to the difference circuit 183.

In practice, adaptive prediction circuits 401 and 402
Applies a class classification adaptive process to predict one pixel in the lower layer from a plurality of upper layer pixels. Specifically, the pixels of the lower hierarchy to be predicted are classified based on the level distribution of a plurality of upper hierarchy pixels spatially nearby. Further, the adaptive prediction circuits 401 and 402 each have a memory (not shown) that stores a plurality of prediction coefficients or one prediction value for each class, which is acquired by learning in advance.
A plurality of prediction coefficients or one prediction value corresponding to the class determined by the class classification is read from the memory. In the case of the prediction value, the prediction value is used as it is as the prediction pixel, and in the case of the prediction coefficient, the prediction value is generated by linear linear combination of the plurality of prediction coefficients and the plurality of pixels.

If the predicted value is normalized,
Predetermined processing is performed on this predicted value to generate a predicted pixel.
For details of such a class classification adaptive process, see
It is disclosed in 4-155719. In addition, such an algorithm for class adaptive processing is already known.
Further, although the classification prediction adaptive algorithm is used in the adaptive prediction circuit of the third embodiment, the present invention is not limited to this, and other prediction methods currently used may be used.

FIG. 9 shows the configuration of an image signal decoding device 190 'which decodes the transmission image data D180' compressed and encoded by the image signal encoding device 400. This image signal decoding device 190 'is different from the image signal decoding device 190 of the second embodiment in that the adaptive prediction circuits 401' and 40 '.
It has the same configuration as the image signal decoding device 190 of the second embodiment except that 2'is provided. Therefore, the parts corresponding to those in FIG. 7 are designated by the same reference numerals.

In the image signal decoding apparatus 190 'according to the third embodiment, the first adaptive prediction circuit 401' to which the third layer decoded data D189 'is input corresponds to the adaptive prediction circuit 401 shown in FIG. Third class adaptation process
This is executed based on the hierarchical decoded data D189 ', and the second hierarchical prediction data D400' obtained as a result is sent to the pixel generation circuit 195. The pixel generation circuit 195 uses the second layer difference / fractional subtraction data D188 'and the second layer prediction data D400' output from the second layer dequantization circuit 192 to execute the thinning circuit 182 shown in FIG. A thinned second layer pixel is generated. Also, the third layer decoded data D
The second layer prediction data D400 ′ obtained by the first adaptive prediction circuit 401 ′ based on 189 ′ is converted into the second layer difference / subtraction data D188 ′ output from the second layer dequantization circuit 192. Is added. The second layer thinned-out decoded data D190 ′ obtained by this addition is combined with the second layer pixel D191 ′ generated by the pixel generation circuit 95 to become the second layer decoded image data D192 ′.

Further, the second adaptive prediction circuit 402 'which receives the second layer decoded data D192' performs the second class classification adaptive processing corresponding to the application prediction circuit 402 shown in FIG.
This is executed based on the hierarchical decoded data D192 ', and the first hierarchical prediction data D401' obtained as a result is sent to the pixel generation circuit 198. The pixel generation circuit 198 uses the first layer difference / fractional subtraction data D187 ′ and the first layer prediction data D401 ′ output from the first layer inverse quantization circuit 191 to execute the thinning circuit 184 shown in FIG. A thinned first layer pixel is generated.

The first layer prediction data D401 'obtained by the second adaptive prediction circuit 402' based on the second layer decoded data D192 'is output from the first layer dequantization circuit 191. First-tier difference subtraction data D1
87 'is added. The first layer thinned-out decoded data D193 'obtained by this addition is combined with the first layer pixel D194' generated by the pixel generation circuit 198,
It becomes the first layer decoded image data D195 '.

According to the configuration of the third embodiment described above, when the lower hierarchy pixel corresponding to the highest hierarchy pixel is quantized according to the polarity of the quantization error when the highest hierarchy pixel is quantized. It is possible to improve the compression efficiency and reduce the image quality deterioration by controlling the quantization characteristic of the so that the quantization error becomes the same as the polarity of the quantization error when the uppermost layer is quantized. In addition, adaptive prediction circuits 401 and 402
The second layer difference data D170 ′ and the first layer difference data D170 ′ obtained from the difference circuits 181 and 183, respectively, by using
Since the residual difference of the hierarchical difference data D171 ′ can be further reduced, the image signal encoding device 400 capable of further reducing the amount of transmission information can be realized.

(4) Fourth Embodiment In the fourth embodiment, as described above in the first to third embodiments, according to the characteristics of the quantization error when the uppermost layer pixel is quantized, Controls the quantization characteristics when quantizing the lower layer pixels that spatially correspond to the upper layer pixels so that the quantization error is the same as the polarity of the quantization error when the upper layer pixel is quantized. In addition to this, the higher the quantization circuit, the larger the number of quantization bits (the smaller the quantization width), and the finer the quantization. By doing so, in addition to the effects of the first to third embodiments, the deterioration of the image quality at the time of decoding can be further reduced.

In the case of the fourth embodiment, in the image signal coding apparatus 140 described above with respect to the first embodiment, the quantizing circuit 147 performs one-bit quantization, and the quantizing circuit 14
Quantization of 4 bits is performed by 8 and 16 is performed by the quantization circuit 149.
Each of the quantizing circuits 147 to 14 so as to quantize the bit.
9 to control the quantization characteristic. As the number of quantization bits in this case, the number of quantization bits is selected in consideration of the degree of influence of the quantization error in the upper layer on the restored pixel value when the lower layer data is restored.

As a result, considering, for example, the above equation (10), the degree of error of the quantization error E (Z ₁₁ ) is
Quantization errors E (X ₁₂ ), E (X ₂₁ ), and E (X ₂₂ ) are 1/16 times larger, and quantization errors E (Y ₁₃ ), E
The degree of (Y ₃₁ ), E (Y ₃₃ ) is the quantization error E (X ₁₂ ),
Since it is 1/4 times as large as E (X ₂₁ ) and E (X ₂₂ ), the error of the quantization error E (Z ₁₁ ) is multiplied by 16 in order to obtain the pixels thinned out at the time of decoding, Quantization error E (Y
₁₃ ), E (Y ₃₁ ), and E (Y ₃₃ ) are multiplied by 4, the error obtained from this is the quantization error E (X ₁₂ ), E (X ₂₁ ), obtained directly from the inverse quantization circuit 165. It is about the same as E (X ₂₂ ). Therefore, the image quality deterioration of the lower layer image based on the quantization distortion in the upper layer can be significantly reduced.

Further, the degree of influence of the quantization error of the upper layer data on the lower layer data is related to the number of layers of layers from the lowest layer and the number of pixels used when generating the upper layer data. Therefore, when determining the number of quantization bits for the upper layer, the quantization that affects the lower layer depends on the number of layers in the layer from the lower layer and the number of pixels used when generating the upper layer data. By setting the number of quantization bits to minimize the error or the number of quantization bits more than that, it is possible to provide an image coding apparatus with reduced image quality deterioration. Although it seems that the amount of information transmitted increases as the number of quantization bits in the upper layer increases, the amount of information transmitted by increasing the number of quantization bits does not increase because the number of pixels transmitted in the upper layer decreases. It is suppressed to a level that does not actually pose a problem.

In the image signal coding apparatuses 180 and 400 described above for the second and third embodiments, the quantizing circuit 185 quantizes 1 bit, and the quantizing circuit 1
The quantization circuit 187 quantizes 2 bits, and the quantization circuit 187 4
Each of the quantizing circuits 185 to 18 so as to quantize the bit.
7 to control the quantization characteristic.

As described above, with respect to the number of quantization bits assigned to each quantization circuit of the first embodiment, the second and third quantization bit numbers are assigned.
In the second and third embodiments, the number of quantization bits assigned to each of the quantizing circuits of the embodiment is reduced.
Difference data between layers is transmitted except for the highest layer data. Therefore, the above equations (3) and (14),
As is clear from a comparison between the equations (7) and (19) and the equations (10) and (22), the multiplication coefficient that depends on the quantization error when restoring the thinned pixels is the first embodiment. This is because the second and third embodiments are suppressed smaller than the above case. That is, the number of quantization bits is determined by the degree of influence of the quantization error on the lower layer.

(5) Other Embodiments In the above embodiment, the case where three layers of hierarchical image data D131, D134, and D135 are generated, compression-encoded and transmitted is described. The present invention is not limited to this, and it is possible to sequentially calculate the average value to obtain 4
The present invention can also be applied to a case where layer image data for layers or 5 layers is generated, and this is compression-encoded by quantization and transmitted.

Further, in the above-described embodiment, the case where one pixel in the upper layer is generated by the average value calculation using the four pixels in the lower layer has been described, but the present invention is not limited to this, and for example, in the lower layer. The hierarchical image data may be generated by generating one pixel in the upper layer by an average value calculation using 6 pixels or more pixels in the layer.

In the fourth embodiment, the number of quantization bits of the quantizing circuits 47, 48 and 49 is set to 1 bit, 4 bits and 16 bits respectively, and the quantizing circuits 85 and 8 are used.
Although the case where the quantization bits of 6 and 87 are 1 bit, 2 bits and 4 bits respectively has been described, the present invention is not limited to this, and the point is that the finer the quantization width is controlled in the quantization circuit of the upper layer. You can do it.

Further, in the above-described embodiments, all quantizers are described as linear quantizers, but the present invention is not limited to this, and nonlinear quantizers, adaptive quantizers, and dynamic range can be used. Appropriate application quantization or the like may be applied. Further, although the embodiment of the present invention is realized by the hardware shown by using the block diagram, the present invention is not limited to this, and it may be realized by software using a CPU, a memory and the like. Is.

Various modifications and application examples can be considered without departing from the gist of the present invention. Therefore, the gist of the present invention is not limited to the embodiments.

[0094]

As described above, according to the present invention, with respect to a plurality of hierarchical image data generated by the average value calculation, a plurality of lower layer pixels used for the same average value calculation to generate one pixel in the upper layer. The pixel data for one pixel among the above is excluded from the transmission target, and the lower layer spatially corresponding to the uppermost layer pixel is spatially corresponding to the characteristic of the quantization error when the pixel data of the uppermost layer is quantized. Improves compression efficiency by controlling the quantization characteristics when quantizing pixel data of layers to be the same as the polarity of the quantization error when quantizing the highest layer pixel data. In addition, it is possible to realize hierarchical coding with reduced image quality deterioration.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an image signal encoding device according to a first embodiment.

FIG. 2 is a schematic diagram used for explaining layering and transmission pixels.

FIG. 3 is a schematic diagram for explaining control of a quantization characteristic.

FIG. 4 is a block diagram showing the configuration of an image signal decoding apparatus according to the first embodiment.

FIG. 5 is a table showing an influence of a quantization error in each layer on a restored pixel in each layer.

FIG. 6 is a block diagram showing a configuration of an image signal encoding device according to a second embodiment.

FIG. 7 is a block diagram showing a configuration of an image signal decoding device according to a second embodiment.

FIG. 8 is a block diagram showing a configuration of an image signal encoding device according to a third embodiment.

FIG. 9 is a block diagram showing a configuration of an image signal decoding device according to a third embodiment.

FIG. 10 is a block diagram showing the configuration of a conventional image signal coding apparatus that realizes hierarchical coding.

FIG. 11 is a block diagram showing the configuration of a conventional image signal decoding apparatus for decoding hierarchically encoded data.

[Explanation of symbols]

140, 180, 400 ... Image signal encoding device, 14
3, 146 ... Averaging circuit, 142, 145, 182,
184 ... Thinning circuit, 147 to 149, 185 to 18
7 ... Quantization circuit, 160, 190 ... Image signal decoding device, 401, 402, 401 ', 402' ... Adaptive prediction circuit, 200 ... Quantization control circuit, D131 ... Input image data, D133 ... ... Second layer image data, D13
5 ... Third layer image data, D141, D177, D1
77 '... First layer encoded data, D142, D17
8, D178 '... Second layer encoded data, D143,
D179, D179 '... Third layer encoded data, D1
44, D180, D180 '... Transmission image data, D1
58, D189 ... Third layer decoded image data, D16
0, D192 ... Second layer decoded image data, D162,
D195 ... First layer decoded image data, D170, D1
70 '... Second layer difference data, D171, D171'
...... First layer difference data, D400 ... Second layer prediction data, D401 ... First layer prediction data, S1 ... Quantization error information signal, S2, S3 ... Quantization control signal, Z ...
... third-tier pixel, Y ... second-tier pixel, X ... first-tier pixel.

Claims (14)

[Claims] 1. An image signal encoding method for producing a plurality of layer image data having different resolutions from input image data and encoding each layer image data, wherein a plurality of pixel values of each layer image data are averaged, For the step of generating the upper layer image data and each layer image data except the uppermost layer having the lowest resolution, among the plurality of lower layer pixels used for the same average value calculation to generate one pixel of the upper layer The step of thinning out pixel data for one pixel to form thinning data formed by the remaining pixels, and compressing the uppermost layer image data and the thinned data of each layer other than the uppermost layer by quantization, respectively. A quantization step of encoding and generating a plurality of layered encoded data; and in the quantization step, the pixel data of the highest layer is quantized. The quantization characteristic when quantizing the pixel data of the lower layer spatially corresponding to the uppermost layer pixel according to the polarity of the quantization error when the quantization error is An image signal coding method characterized by controlling so that it has the same polarity as the quantization error when quantized. 2. The image signal coding method according to claim 1, wherein in the quantization processing step, data in a higher layer is quantized more finely. 3. An image signal encoding method for generating a plurality of layer image data having different resolutions from input image data and encoding each layer image data, averaging a plurality of pixel values of each layer image data, For the step of generating upper layer image data and the step of forming the layer difference data by calculating the difference between each layer image data and the adjacent upper layer image data for each layer image data except the highest layer data having the lowest resolution. With respect to the layer difference data of each layer, the difference pixel data that can be restored by the arithmetic operation using the difference data of the adjacent upper layer and the difference data of the own layer is decimated, and the decimated data formed by the remaining pixels is extracted. The steps to be formed and the above-mentioned top-level image data and the thinned-out data of each layer other than the above-mentioned top layer are respectively quantized. And compression coding, and a quantization step for generating a plurality of hierarchically encoded data, and, in the quantization step, depending on the polarity of the quantization error when the pixel data of the highest layer is quantized, The quantization characteristic when quantizing the pixel data of the lower layer corresponding to the pixel of the highest layer is set so that the quantization error becomes the same as the polarity of the quantization error when the pixel data of the highest layer is quantized. An image signal encoding method, characterized in that the image signal encoding method is characterized in that: 4. The image signal coding method according to claim 3, wherein in the quantization step, data in a higher hierarchy is quantized more finely. 5. The step of calculating the layer difference data according to claim 3, wherein an adjacent lower layer predicted and generated by a predetermined prediction calculation process using each of the layer image data and the adjacent upper layer image data. An image signal encoding method, wherein the hierarchical difference data is calculated by taking a difference from image data. 6. An image signal encoding apparatus for generating a plurality of layer image data having different resolutions from input image data and encoding each layer image data, averaging a plurality of pixel values of each layer image data, For the unit that generates the upper layer image data and each layer image data except the uppermost layer with the lowest resolution, among the plurality of lower layer pixels used for the same average value calculation to generate one pixel in the upper layer A unit for thinning out pixel data for one pixel to form thinning data formed by the remaining pixels, and the above-mentioned highest layer image data and the thinning data for each layer other than the uppermost layer are respectively compressed by quantization. And a quantization unit that generates a plurality of layered encoded data, and the above quantization unit quantizes the pixel data of the highest layer. The quantization characteristic when quantizing the pixel data of the lower layer spatially corresponding to the uppermost layer pixel according to the polarity of the quantization error when the quantization error is An image signal encoding device, which is controlled so as to have the same polarity as a quantization error when quantized. 7. The image signal coding apparatus according to claim 6, wherein the quantization unit is further quantized more finely as the data in the higher hierarchy. 8. An image signal encoding device for generating a plurality of layer image data having different resolutions from input image data and encoding each layer image data, averaging a plurality of pixel values of each layer image data, A unit that generates upper layer image data and a unit that calculates the difference between each layer image data and the adjacent upper layer image data for each layer image data except the highest layer data with the lowest resolution, and forms the layer difference data. With respect to the layer difference data of each layer, the difference pixel data that can be restored by the arithmetic operation using the difference data of the adjacent upper layer and the difference data of the own layer is decimated, and the decimated data formed by the remaining pixels is extracted. The unit to be formed and the above-mentioned top-level image data and the thinned-out data of each layer except the above-mentioned top layer are respectively quantized. Compression coding, and a quantization unit for generating a plurality of hierarchically encoded data, and, the quantization unit, depending on the polarity of the quantization error when the pixel data of the highest layer is quantized, The quantization characteristic when quantizing the pixel data of the lower layer corresponding to the pixel of the highest layer is set so that the quantization error becomes the same as the polarity of the quantization error when the pixel data of the highest layer is quantized. An image signal coding device characterized in that it is controlled to. 9. The image signal coding apparatus according to claim 8, wherein the quantization unit is such that data of higher layers is quantized more finely. 10. The unit for calculating the layer difference data according to claim 8, wherein the adjacent lower layer predicted and generated by a predetermined prediction calculation process using each layer image data and the adjacent upper layer image data. An image signal encoding device, wherein the hierarchical difference data is calculated by taking a difference from image data. 11. An image signal transmission method for generating a plurality of layer image data having different resolutions from input image data, encoding each layer image data, and transmitting the encoded data, wherein a plurality of pixels of each layer image data are transmitted. A step of averaging the values to generate upper layer image data, and a plurality of layers used for the same average value calculation to generate one pixel of the upper layer for each layer image data except the highest layer having the lowest resolution. The step of thinning out pixel data for one pixel of the lower layer pixels and forming thinning data formed by the remaining pixels, and the thinning data of each layer except the uppermost layer image data and the uppermost layer. A quantization step for compression-encoding by quantization to generate a plurality of hierarchically encoded data as transmission image data, and transmitting the transmission image data. In the quantization step, the pixel of the lower layer spatially corresponding to the uppermost layer pixel is spatially corresponding to the polarity of the quantization error when the pixel data of the uppermost layer is quantized. An image signal transmission method, characterized in that the quantization characteristic when quantizing data is controlled so that the quantization error becomes the same as the polarity of the quantization error when quantizing the uppermost layer pixel data. . 12. An image signal transmission method for generating a plurality of layer image data having different resolutions from input image data, encoding each layer image data, and transmitting the encoded data, wherein a plurality of pixels of each layer image data are provided. The step of averaging the values to generate upper layer image data and the difference between each layer image data and the adjacent upper layer image data is calculated for each layer image data except the highest layer data with the lowest resolution With respect to the step of forming the difference data and the layer difference data of each layer, the difference pixel data which can be restored by the arithmetic operation using the difference data of the adjacent upper layer and the difference data of the own layer is thinned, and the remaining pixels are The step of forming the thinned-out data to be formed, and the thinning-out data of each layer other than the uppermost layer image data and the uppermost layer. Each of which is compression-encoded by quantization to generate a plurality of layer-encoded data as transmission image data, and a step of transmitting the transmission image data, and the quantization step, Depending on the polarity of the quantization error when the pixel data of the highest layer is quantized, the quantization error when quantizing the pixel data of the lower layer pixel data corresponding to the highest layer pixel is An image signal transmission method, characterized in that the uppermost layer pixel data is controlled so as to have the same polarity as that of a quantization error when the pixel data is quantized. 13. A recording medium that can be decoded by a decoding device, wherein the recording medium has a recording signal that can be decoded by the decoding device, and the recording signal is a plurality of pixel values of each hierarchical image data. Is calculated to generate upper layer image data, and for each layer image data except the uppermost layer having the lowest resolution, a plurality of lower layers used for the same average value calculation to generate one pixel in the upper layer. The step of thinning out pixel data for one pixel of the hierarchical pixels to form thinning data formed by the remaining pixels, and the thinning data for each layer except the uppermost layer image data and the uppermost layer are respectively quantized. And a quantization step for generating a plurality of hierarchically coded data by compression coding, and in the quantizing step, the pixel data of the uppermost hierarchy is According to the polarity of the quantization error when quantized, the quantization characteristics when quantizing the pixel data of the lower layer spatially corresponding to the uppermost layer pixel, the quantization error is the highest layer pixel A recording medium which is controlled so as to have the same polarity as the quantization error when the data is quantized. 14. A recording medium that can be decoded by a decoding device, wherein the recording medium has a recording signal that can be decoded by the decoding device, and the recording signal is a plurality of pixel values of each hierarchical image data. For each layer image data except the highest layer data with the lowest resolution, the difference between each layer image data and the adjacent upper layer image data is calculated, and the layer difference is calculated. With respect to the step of forming data, the difference pixel data of each of the above layers is thinned out by subtracting difference pixel data which can be restored by an arithmetic operation using the difference data of the adjacent upper layer and the difference data of its own layer, and forms the remaining pixel. The step of forming the thinned-out data to be quantized, and the thinned-out data of each layer other than the uppermost layer image data and the uppermost layer are respectively quantized. Therefore, it is formed from a quantization step for compression-encoding and generating a plurality of layer-encoded data, and a step for transmitting the transmission image data. In the quantization step, the pixel data of the highest layer is quantized. Depending on the polarity of the quantization error at that time, the quantization characteristic when quantizing the pixel data of the lower layer corresponding to the uppermost layer pixel, the quantization error quantized the uppermost layer pixel data A recording medium, which is controlled so as to have the same polarity as the quantization error.