JPH09307911A - Image coder, image decoder and method therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To attain resolution conversion with less image quality deterioration. SOLUTION: An image signal Si representing a pixel value and a significance signal Ss denoting whether or not a pixel value is significant of each pixel are received at first as input signals. Then the vicinity of a pixel element of the significance signal Ss to be received is referenced, and a resolution conversion characteristic decision device 103 selects a prescribed resolution conversion characteristic among two kinds of resolution conversion characteristics or over used to conduct resolution conversion of the image signal Si by using the significant pixels only. A resolution converter 105 used the selected resolution conversion characteristic selected as above to apply resolution conversion to the input image signal Si and provides an output of the result as an image converter output signal Sr. Through the processing above, the resolution conversion such as pixel thinning and pixel interpolation without being affected by insignificant pixels in the received image signal Si is conducted and the resolution conversion with less deterioration through the conversion is realized.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for encoding and decoding an image signal used for efficiently transmitting or storing the image signal, and a method thereof.

[0002]

2. Description of the Related Art In order to efficiently transmit or store an image signal representing an image of an object as a group of pixels, especially a moving image signal, a method of coding the moving image by dividing it into different layers for each object is proposed. Has been done. For example, the encoding of an image composed of a person and a background will be described as an example. First, the image is separated into two layers of a person and a background, and encoding and transmission are performed for each layer.

On the other hand, in the image decoding device, the transmitted signal encoded for each layer is decoded, and each decoded image is combined into a single layer image by a predetermined method and displayed. To do. In addition, at the time of composition, it is necessary to provide information indicating, for each pixel, whether or not the background is hidden by the superimposed images. This information used for compositing is called significant information, and pixels whose background is hidden are called significant.
Further, the significant information can provide the correlation information between pixels even when the significant information is encoded as one image without being separated into a plurality of layers. That is, significant means that the pixel is located inside the object, and insignificant means that the pixel is located outside the object. A large value of significant information means that the combined ratio is large and is visually important. Conversely, a small value of significant information means that it is rarely used in composition, that is, the transparency is high. Represent

As described above, a signal composed of significant information corresponding to an arbitrary pixel in the image signal is called a significant signal. When superimposing a plurality of images composed of image signals, it is possible to represent by a significant signal whether or not the background is hidden by the superimposed images. In that case, the non-zero value of the significant signal indicates that the pixel hides the background pixel. Further, “not significant” means that there is no significant signal value, and the pixel is transparent, so that it is not used in synthesis. That is, the significant signal represents the shape of the image to be combined with the background, and only the significant pixels affect the image quality of the combined image. In other words, the insignificant pixels are irrelevant to the image quality, and the encoding efficiency can be improved by encoding only the significant pixels.

[0005]

However, in the conventional image signal encoding apparatus, when encoding an image signal having a large amount of pixel value information such as luminance, color difference, and transparency for each pixel, pixel thinning or pixel When the interpolation is performed, the same frequency conversion is performed regardless of whether or not the pixel is significant. Therefore, the pixel value of an insignificant pixel, that is, a meaningless pixel, affects a significant pixel, thereby deteriorating the image quality.

Further, although the resolution of the luminance signal of the image signal and the resolution of the significant signal are the same, the resolution of the color difference signal may be different from the resolution of the significant signal. In such a case, there is no problem in encoding the luminance signal based on the significant signal.
However, in order to encode the color difference signal based on the significant signal, it is necessary to make the resolutions of both signals the same. That is,
Prior to encoding the color difference signal, it is necessary to convert the resolution of the significant signal to be the same as the resolution of the color difference signal. However, as described above, even in the resolution conversion of the significant signal, the same resolution conversion is performed regardless of whether or not the target pixel is significant. Had deteriorated.

Further, in the conventional image signal encoding device and image signal decoding device, a pixel which is almost transparent and has no visual effect due to a minute noise component of the input signal or encoding distortion is treated as a significant pixel. Therefore, the pixel value of the pixel which is almost transparent is also required to be encoded, which lowers the encoding efficiency.

In view of the above points, the present invention realizes a sharp converted image by performing resolution conversion using only significant pixels, and further makes pixels that are nearly transparent into non-effective pixels to be encoded. By reducing the number of bits to reduce the number of coding bits, the image signal can be efficiently coded without degrading the image quality, and the coded image signal can be efficiently decoded to reproduce a high quality image. It is an object of the present invention to provide an image signal encoding / decoding device and method capable of performing the same.

[0009]

As a means for solving the above-mentioned problems, the present invention comprises a significant element that represents a significant value of each pixel of an image signal composed of two or more adjacent pixels. In the image signal coding device for coding the image signal based on the significant signal, the resolution conversion characteristic determining means for determining the resolution conversion characteristic based on the significant value of the pixel in the vicinity of the pixel, and the determined resolution conversion characteristic. A resolution conversion means for converting the resolution of the image signal according to the resolution characteristic is provided.

[0010]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention will be described below with reference to specific embodiments. In the description of this embodiment,
The input image signal will be described by taking an image signal composed of a two-dimensional color signal representing a pixel value and a significant signal indicating whether or not the pixel value is significant for each pixel of the color signal as an example. Image information other than the color signal may be included, or the image signal itself may be a signal representing image information of any dimension other than two dimensions.

(Embodiment 1) First, referring to FIG.
The structure of the image signal encoding device EC1 according to the present invention will be described. The image signal encoding device EC1 includes a first input terminal Ti1, a second input terminal Ti2, a resolution conversion characteristic determiner 103, a resolution converter 105, and an output terminal To. The first input terminal Ti1 is connected to an external image signal source (not shown) and receives the supply of the image signal Si.
Similarly, the second input terminal Ti2 is connected to an external significant signal source (not shown) and receives the significant signal Ss having significant information corresponding to each pixel in the image signal Si. The significant signal Ss is a multilevel signal (Multilevel signa) that indicates the mutual relationship between pixels in the image signal Si with two or more values.
However, in the present embodiment, for simplification of description, a case of a binary signal (Two Level signal) indicating two states of whether or not the significant signal Ss exists will be mainly described. explain. For example, when the significant signal Ss is binary, it represents information indicating whether or not the foreground pixel is transparent, but when the significant signal Ss is multivalued, the pixel indicating how many background pixels the foreground pixel displays. It is possible to express the transparency between them. Further, when the image is encoded, by controlling the encoding of the target pixel based on the significant information of the target pixel and the surrounding pixels,
Efficient encoding processing is possible.

The resolution conversion characteristic determiner 103 is connected to the second input terminal Ti2 and receives the significant signal Ss.
The resolution conversion characteristic determiner 103 receives the input significant signal Ss.
, The resolution conversion characteristic optimum for performing resolution conversion of the image signal Si signal is determined by using only significant pixels, and the resolution conversion characteristic instruction signal SL indicating the determined resolution conversion characteristic is determined. To generate. The method of determining the resolution characteristic by the resolution conversion characteristic determiner 103 will be described later in detail with reference to FIGS. 3 and 4.

The resolution converter 105 has a first input terminal Ti.
1 and the resolution conversion characteristic determiner 103 to be supplied with the image signal Si and the significant signal Ss, respectively. The resolution converter 105 previously stores therein the resolution conversion characteristic data corresponding to the respective states of the significant pixels of the significant signal Ss, and obtains the resolution characteristic data corresponding to the resolution characteristic indicated by the resolution conversion characteristic instruction signal SL. Using the image signal S
Resolution conversion such as pixel thinning out of i and pixel interpolation is performed, and the converted resolution image signal Si is output from the output terminal To.
In this way, since resolution conversion processing such as pixel thinning or pixel interpolation that is not affected by insignificant pixels can be performed on the input image signal Si, it is possible to prevent deterioration of image quality due to resolution conversion processing. The resolution conversion processing of the resolution converter 105 will be described later in detail with reference to FIGS. 5, 6, 7, 8, 9, and 10.

In this example, the resolution conversion of the image signal Si is described based on the significant signal Ss.
By inputting the significant signal Ss to the resolution converter 105 instead of the image signal Si at
It goes without saying that the resolution conversion of the significant signal Ss can be performed based on the significant signal Ss. Further, such resolution conversion of the significant signal Ss will be described later with reference to FIG.

The main operation of the image signal encoding of the image signal encoding device EC1 will be described with reference to the flowchart shown in FIG. When the image signal encoding operation is started,
First, an image signal Si and a significant signal Ss are generated by an external image signal source and a significant signal source, respectively.

At step # 100, the first input terminal Ti1
The image signal Si is input to the resolution converter 105 through the, and the significant signal Ss is input to the resolution conversion characteristic determiner 103 through the second input terminal Ti2.

In step # 200, the resolution conversion characteristic determiner 103 inputs the significant signal Ss, that is, the image signal S.
It is determined whether all the significant information in the significant signal Ss corresponding to each pixel of i is significant. In this example, the significant signal Ss
Is a binary signal, the pixel is significant if the value indicating the significant information for each pixel in the significant signal Ss is not zero, while the pixel is not significant if the significant information is zero. When all the significant information in the significant signal Ss is significant, YES is determined and the resolution conversion characteristic determiner 103 is determined.
Generates a resolution conversion characteristic selection signal SL indicating a normal resolution conversion characteristic, and proceeds to the next step # 300. An example of a normal resolution conversion characteristic will be described later with reference to FIG.

In step # 300, the resolution converter 10
5 selects a normal resolution characteristic from the resolution conversion characteristic data stored inside based on the resolution conversion characteristic selection signal SL generated in step # 200, and then proceeds to step # 500.

On the other hand, if NO in step # 200, that is, if it is determined that at least one of all the significant information in the significant signal Ss is not significant, the resolution conversion characteristic determiner 103 determines the significant pixel. After the resolution conversion characteristic designating signal SL for designating the resolution conversion characteristic corresponding to the state of the significant pixels in the current significant signal Ss is generated so that the resolution conversion is performed by using only, the process proceeds to the next step # 400.

In step # 400, the resolution converter 10
5 selects a resolution characteristic corresponding to the current significant signal state from the resolution conversion characteristic data stored inside based on the resolution conversion characteristic selection signal SL, and then proceeds to step # 500.

In step # 500, step # 300
And on the basis of the resolution characteristics selected in step # 400, the image signal Si is subjected to resolution conversion to generate the resolution-converted image signal Sr, and then the next step # 6.
Go to 00.

In step # 600, the resolution converter 105
After the resolution-converted image signal Sr generated by is output from the output terminal To to the outside, the processing is ended.

As described above, since the resolution conversion of the image signal Si is performed based on the significant signal Ss, the resolution conversion processing can be further performed without the influence of insignificant pixels of the input image signal Si. It is possible to prevent deterioration and realize efficient image signal coding. Note that step #
In 400, the resolution conversion can be performed not only on the image signal Si but also on the significant signal Ss itself by performing the resolution conversion on the significant signal Ss.

Next, the resolution conversion operation of the resolution converter 105 shown in FIG. 1 will be described with reference to FIGS. 3 and 4. In FIG. 3, the two-dimensional color signal Sc in the image signal Si
And the significant signal Ss corresponding to the same color signal Sc. P0 in the color signal Sc is the coordinate (X,
Y) shows a color pixel Sc (P0),
1 is a color pixel Sc adjacent to the color pixel Sc (P0)
(P1) is shown. Further, α0 and α1 in the two-dimensional significant signal Ss indicate significant pixels Ss (α0) and Ss (α1) in the significant signal Ss corresponding to the pixels P0 and P1 described above, respectively.

FIG. 4 shows significant pixels Ss (α0) and Ss.
An example of the resolution conversion characteristic function for the value of (α1) will be shown.
In the figure, the first column shows significant pixels Ss (α0) and Ss.
The combinations of the values of (α1) are defined as conditions C1, C2, C3, and C.
It is classified as 4. The second and third columns show the values of the significant pixels Ss (α0) and Ss (α1), respectively, where 1 is significant and 0 is not significant, as described above. is there. Significant pixel S in the first column
The resolution conversion characteristic Fc corresponding to the values of s (α0) and Ss (α1) is represented as a function Pf = fα0, α1 (P0, P1). In the same column, P0 and P1 represent the pixel values of the color pixels Sc (P0) and Sc (P0), respectively.

Under the condition C1, both significant pixel values Ss (α0)
And Ss (α1) are both 1, that is, both color pixels Sc
If (P0) and SC (P1) are both significant,
The resolution conversion characteristic function can be expressed by the following equation.

[0027]

## EQU1 ## Pf = (P0 + P1) // 2

That is, the average of both pixel values is used as the resolution characteristic function. In addition, this Expression 1 is step # 300.
It corresponds to the normal resolution conversion characteristic selected in.

Under the condition C2, the significant pixel value Ss (α0) is 1, while the significant pixel value Ss (α1) is 0, that is,
When only the color pixel Sc (P0) is significant, the resolution conversion characteristic function can be expressed by the following equation.

[0030]

(2) Pf = P0

That is, the color pixel Sc (P0) obscures the color pixel Sc (P1).

Under the condition C3, unlike the condition 2, when only the color pixel Sc (P1) is significant, the resolution conversion characteristic function can be expressed by the following equation.

[0033]

[Formula 3] Pf = P1

That is, the color pixel Sc (P0) is covered by the color pixel Sc (P1).

Under the condition C4, both significant pixel values Ss (α0)
And Ss (α1) are both 0, that is, both color pixels Sc
If both (P0) and SC (P1) are not significant,
The resolution conversion characteristic function Pf is not generated.

Next, the operation of the resolution converter 105 will be described with reference to FIGS. 5, 6, 7, 8, 9 and 10.

FIG. 5 schematically shows the conversion of 2 × 2 pixels of the image signal Si into one pixel of the resolution-converted image signal Sr, that is, resolution conversion for reducing the image to 1/4. That is, the image signal Si having 4 × 4 pixels is 2 × 2
The image signal Sr of the pixel is converted. Here, a method of converting the 4-pixel portion of the image signal Si, for example, the image signal block Sib composed of the lower left 2 × 2 pixels to the lower left 1 pixel K of the resolution-converted image signal Sr will be described.

In the figure, the image signal Si is composed of 4 × 4 pixels, and the significant signal Ss corresponding to the image signal Si is also composed of 4 × 4 significant elements. Each significant element has a significant information value regarding the corresponding pixel of the image signal Si. The image signal block Sib corresponds to the significant signal block Ssb composed of the lower left 2 × 2 significant elements in the significant signal Ss. In the image signal Si, the image signal block Sib, the significant signal Ss, and the significant signal block Ssb, the significant pixels and elements are shaded and displayed. Specifically, the pixels W, X, Y, and Z of the image signal block Sib correspond to the significant elements A, B, and D in the significant signal block Ssb, respectively, and the significant pixels in the image signal block Sib are , The image signal block Sib has three pixels of W, X, and Z. Pixels W, X, which are significant for insignificant pixel Y,
Since a value having no correlation with Z may be included, if the resolution conversion is performed using the pixel values of all the pixels in the image signal block Sib, the image signal Sr after the resolution conversion is not affected by the insignificant pixel Y. I will receive it.

Therefore, if the value I obtained by averaging the pixel values of the significant pixels W, X, and Z in the image signal block Sib is taken as the value after resolution conversion, the resolution conversion is performed without being affected by the insignificant pixels. It can be carried out. Now pixel W,
If the pixel values of X and Z are Pw, Px, and Pz, respectively,
The average pixel value can be expressed by the following equation.

[0040]

[Equation 4] I = (Pw + Px + Pz) // 3

As described above, the pixel K of the resolution-converted image signal Sr can be obtained by performing the resolution conversion on the four pixels of the image signal block Sib of the image signal Si with the resolution conversion characteristic defined by the equation (4). it can. 2x2 now
An example of converting a pixel into one pixel has been described, but N × M
The resolution conversion of an image signal block having a size (N and M are natural numbers) at an arbitrary reduction ratio can be performed in the same manner as described above. Although the case where the resolution of the image signal is converted has been described here, it has already been described that the resolution of the significant signal can also be converted by the same method. A specific method thereof will be described below with reference to FIG. To do.

In FIG. 6, the significant signal Ss is represented as a signal composed of 4 × 4 significant elements. As in the case of FIG. 5, the 1/4 reduction processing of the significant signal block Ssb of 2 × 2 significant elements in the lower left part of the significant signal Ss into the significant signal SsR of one pixel will be described. A value J obtained by averaging the significant signal values of the significant pixels A, B, and D in the significant signal block Ssb is set as the value after resolution conversion.
The significant values of the significant pixels A, B, D are respectively Va, Vb,
If Vd is set, the average significant value can be expressed by the following equation.

[0043]

[Equation 5] J = (Va + Vb + Vd) // 3

With reference to FIGS. 5 and 6, the resolution conversion method is represented by equations (4) and (5) by taking an example of resolution conversion of 2 × 2 pixels.
However, assuming that the number of pixels of the conversion source is an arbitrary positive number n, when the significant signal value is binary, the pixel value p of the converted image signal Sir can be expressed by the following equation.

[0045]

(Equation 6)

P0 to pn shown on the right side of Expression 6 can be expressed by respective pixel values of the image signal Si to be resolution-converted, and α can be expressed by the value of the significant element of the referenced significant signal Ss. When the significant signal value is multi-valued, the pixel value p of the converted image signal Sir can be expressed by the following equation.

[0047]

(Equation 7)

P0 to pn shown on the right side of Expression 7 are pixel values of the image signal Si to be resolution-converted, w0.
From wn, the value of each significant element of the significant signal Ss is 0.
It can be represented by a pointer indicating whether or not. For example,
When the value of the significant element is 0, the pointer w has a value of 0, and when the value of the significant element is not 0, the pointer w has a value of 1. P on the left side is the pixel value of the image signal Sir converted by the expression on the right side. In the above two examples of resolution conversion, the case of taking an average has been described, but when the target input is binary, the logic of each significant element value of the significant signal Ss is expressed as shown in the following formula 8. Resolution conversion can be realized by taking the sum.

[0049]

[Expression 8] α = α0 [+] α1 [+] ・ ・ ・ [+] αn-1

In Expression 8, [+] means a logical sum. As shown, by taking the logical sum of the respective significant element values, the transformation α becomes significant if even one of the plurality of significant elements is significant. On the other hand, if even one of the plurality of significant elements is not significant, α may be insignificant by constructing the logical product [x] instead of the logical sum [+] as in the following equation. .

[0051]

[Formula 9] α = α0 [×] α1 [×] ・ ・ ・ [×] αn-1

Next, referring to the flowchart of FIG.
The image signal encoding operation of the image signal encoding device EC1 briefly described with reference to the flowchart of FIG. 2 will be described in more detail with respect to the resolution conversion operation of the significant signal Ss having N significant elements. In this example, a method of converting the resolution of the significant signal Ss itself will be described, but it can be applied to the resolution conversion of the image signal Si as described above.

In step S2, the system for resetting the significant counter CS, the significant position pointer PS, the insignificant counter CNS, the insignificant position pointer PNS, and the significant element counter n is initialized, and then the process proceeds to step S4. In the present embodiment, the significant counter CS and the insignificant counter CN
S and the significant element counter n are each set to 1, and the significant position pointer PS and the insignificant position pointer PNS are cleared.

In step S4, the resolution conversion characteristic determiner 103 reads the nth significant element Ssn in the significant signal Ss to be converted every one significant state detection cycle, and then proceeds to the next step S6. In addition, n is the above-mentioned significant element counter n, and the significant element whose significant state is detected first is represented as Ss1. The value of the significance counter CS is also 1.

In step S6, it is determined whether the value of the read significant element Ssn is 0, that is, whether the significant element is significant. When the significant element value is not zero, that is, when the target significant element is significant, YES is determined and the process proceeds to the next step S8.

In step S8, the position information of the current significant element Ssn in the target significant signal Ss is recorded in the significant element position pointer PS based on the current value of the significant element counter n and the value of the significant counter CS, Next step S10
Proceed to. The values of the significant element counter n and the significant element counter CS are both 1 when the first significant element is detected after the resolution conversion process is started.

In step S10, the significant counter CS and the significant element counter n are each incremented by 1, and then the process proceeds to the next step S16.

On the other hand, if NO in step S6, that is, if it is determined that the current target significant element is not significant, step S6.
Proceed to 12. Based on the current value of the significant element counter n and the value of the insignificant counter CNS, the position information of the current significant element Ssn in the target significant signal Ss is set to the significant element position pointer P.
After recording in S, the process proceeds to the next step S14. When the first significant element is detected after the resolution conversion process is started, the values of the significant element counter n and the insignificant counter CNS are both 1.

In step S14, the non-significance counter CN
After S and the significant element counter n are each incremented by 1, the process proceeds to the next step S16.

In step S16, it is determined whether there is any significant element for which no significant state has been detected. That is, the number of significant elements in the significant signal Ss of the significant state detection target is Nm.
If it is ax, it is determined whether n is smaller than Nmax. In the case of YES, that is, it means that there remains a significant element for which a significant state has not been detected yet.
It returns to step S2. On the other hand, if NO in step S16, that is, if it is determined that the detection of the significant state of the detection-target significant element is completed, the process proceeds to next step S18.

The path from step S4 to step S16 described above is called one significant state detection cycle. In this way, the significant state detection cycle is repeated until all the significant elements in the target significant signal Ss have significant states detected until NO is determined in step S16.

In step S18, the previous step S8,
Alternatively, based on the significant element information detected in step S12, the resolution conversion characteristic determiner 103 determines the optimum resolution conversion characteristic for the current significant signal Ss, generates the resolution conversion characteristic instruction signal SL, and then proceeds to step S20. move on.

In step S20, a predetermined resolution conversion characteristic is read from the inside based on the resolution conversion characteristic instruction signal SL generated in step S18, and then the process proceeds to the next step S22.

In step S22, the significant signal Ss is resolution-converted based on the resolution conversion characteristic read in step S20 to generate the resolution-converted image signal Sr, and the post-processing is terminated.

In the flow chart shown in FIG. 7, as a result of the judgment as to whether or not the target significant element in step 6 is significant, steps S8 and S1 are performed for elements that are significant.
The significant state is detected at 0, and the significant state is detected at steps S12 and S14 for insignificant elements. However, regardless of whether the significant element is significant or insignificant, its position and value can be specified by the value of the significant element counter. That is, after step S6,
Even if the significant information is obtained for the significant elements only in steps S8 and S10, the information about the non-significant elements can be obtained. In such a case, step S12
Alternatively, steps S14 and S14 may be omitted, and if NO is determined in step S6, the process may directly proceed to step S16. Similarly, it goes without saying that steps S8 and S10 may be omitted and significant information may be obtained based on non-significant elements in steps S12 and S14.

FIG. 8 illustrates a further example of the resolution conversion method according to the present invention. In the same figure, steps S12 and S14 from the flowchart shown in FIG.
16 and returns from step S16 to step S4, and from step S10 directly to step S4.
It was designed to proceed to 18. As a result, as in the flowchart shown in FIG. 7, it is not necessary to wait until the detection of significant values for all significant elements is completed, and if one of the target significant elements is significant, the resolution conversion process is performed. It is configured in. Further, it goes without saying that by replacing Steps S12 and S14 with Steps S8 and S10, if even one of the target significant elements is insignificant, the resolution conversion process can be performed.

Next, the enlarged pixel conversion will be described with reference to FIG. In the figure, the image signal Si of 3 pixels is set to 5
The resolution conversion for converting the pixel resolution conversion image signal Sr is schematically shown. The image signal Si is composed of three pixels A1, A2 and A3, and the significant signal Ss corresponding to the image signal Si is also composed of three elements X1, X2 and X3. Each of the pixels A1, A2, and A3 corresponds to each of the elements X1, X2, and X3, and the pixels X1 and X2 are significant. The pixels X1, X2, and X3 located at odd-numbered positions from the resolution conversion image signal Sr correspond to the pixels X1, X2, and X3 of the image signal Si as they are, but the pixels Y1 located at even-numbered positions from the top. Regarding Y2 and Y2, referring to the image signal Si,
The resolution conversion is performed only for the pixels that are significant. That is, the pixels Y1 and Y2 are converted from the significant elements X1 and X2, and the conversion characteristic can be expressed by the following equation.

[0068]

## EQU10 ## Y1 = (X1 + X2) / 2 Y2 = X2

X1, X2, Y1, and Y2 in the equation 10 are pixel values of corresponding pixels.

Next, referring to FIG. 10, the resolution conversion of the two images Ia and Ib which are temporally continuous will be described. As shown in the figure, the image Ia that is ahead in time is shown.
Is smaller than the subsequent image Ib in terms of time, by performing the enlargement resolution conversion described with reference to FIG. 9 between the preceding image signal Sia and the subsequent image signal Sib corresponding to each image. More efficient correspondence can be performed. For example, it can be applied to scalability.

(Embodiment 2) Next, with reference to FIG. 11, an image signal coding apparatus EC2 according to a second embodiment of the present invention will be described. Main image signal encoding device EC
2, the image signal encoding device E in the first embodiment
Unlike C1, the resolution conversion is performed only on the significant pixels of the image signal Si. Image signal encoding device EC2
Is similar to the structure of the image signal encoding device EC1 shown in FIG. 1, but performs resolution conversion of the significant signal Ss between the resolution conversion characteristic determiner 103 and the second input terminal Ti2 to obtain the resolution converted significant signal Ssr. Second resolution converter 3 for generating
01 is added to the configuration. Further, the resolution conversion significant signal Ssr is output to the resolution conversion characteristic determiner 103.
It is also output to the output terminal To.

The second resolution converter 301 receives the input Ss in the same manner as described for the resolution converter 105.
Image thinning or image interpolation is performed. The resolution conversion characteristic determiner 103 refers to the significant signal in the vicinity of the pixel of the resolution conversion significant signal Ssr, and determines the resolution conversion characteristic such that the resolution conversion of the image signal Si is performed using only the significant pixels. The resolution conversion characteristic instruction signal SLa to be selected is generated. When the resolution conversion significant signal Ssr is not significant, the conversion of the image signal Si for deriving the pixel is not necessary, and therefore the resolution conversion characteristic instruction signal SLa for instructing the resolution characteristic represented by the function that is the simplest to calculate. Is generated. The resolution converter 105 uses the image signal encoding device EC1.
Similarly to the above, the image signal Si is resolution-converted with the resolution conversion characteristic designated by the resolution conversion characteristic designation signal SLa, and is output to the output terminal To as the resolution-converted image signal Sra.

By the above processing, only the significant pixels whose resolution is converted by the second resolution converter 301 are similar to those executed by the image signal encoding device EC1 according to the first embodiment. By performing resolution conversion such as pixel thinning or pixel interpolation that is not affected by the insignificant pixel of, the equivalent conversion image quality can be realized with a smaller number of function calculations.

(Third Embodiment) Next, referring to FIG. 12, an image signal encoding device EC3 according to the third embodiment of the present invention.
Will be described. The image signal encoding device EC3 is an image encoding device that performs an encoding process on the significant signal Ss after performing an N-value encoding process by a threshold value processor. The image signal coding device EC3 also has a structure similar to that of the image signal coding device EC1 shown in FIG.
It has a third input terminal Ti3 for receiving the input of the threshold signal Sth having h. Furthermore, a threshold value processor 402 is added between the second input terminal Ti2 and the resolution conversion characteristic determiner 103. The threshold processor 402 has a third input terminal Ti.
3 is also connected to receive the significant signal Ss and the threshold signal Sth, respectively.

Further, the resolution converter 105 and the output terminal To
An encoder 405 is added between and. The encoder 405 is further connected to the third input terminal Ti3 and supplied with the threshold signal Sth.

The threshold value processor 402 compares the significant signal Ss with the threshold value signal Sth. If the significant value of the significant signal Ss is smaller than the threshold value signal Sth, the significant value of the significant signal Ss is set to an insignificant value. Converted and threshold-converted significant signal Ss
It is output to the resolution conversion characteristic determiner 103 as t. still,
The significant signal Ss having an insignificant value is not processed by the threshold value processor 402 and is output as it is as the significant signal Sst.

The resolution conversion characteristic determiner 103 outputs a resolution conversion characteristic instruction signal SLb for instructing the optimum resolution characteristic for the image signal Si to the resolution converter 105 based on the threshold conversion significant signal Sst. The resolution converter 105 converts the image signal Si supplied from the first input terminal Ti1 with an optimum resolution conversion characteristic based on the resolution conversion characteristic instruction signal SLt to generate a resolution converted image signal Srb.

The encoder 405 encodes the resolution-converted image signal Srb supplied from the resolution converter 105 and the threshold signal Sth supplied from the third input terminal Ti3,
The encoded image signal Sic1 is output from the output terminal To.

Referring to FIG. 13, threshold conversion significant signal Ss
The significance of generating t will be briefly described. In the present invention, as described above, the image signal S
i or the significant signal Ss itself is encoded, but when the significant signal Ss is multi-valued, even a low-level significant value that is nearly transparent and visually indistinguishable from other pixels is encoded. This will adversely affect the chemical conversion process. An example of such a significant signal Ss is shown in the curve B of FIG. This significant signal Ss
(B) has a value greater than zero in the area indicated by the arrow As, and is therefore significant over the area As.

However, both ends A1 and A of the region As are
2 is a factor that causes deterioration of encoded data, although the significance is low so that it cannot be visually distinguished from other areas as described above. Therefore, in the region As of the significant signal Ss, the significant signal Ss shown by the curve A in FIG. 13 can be obtained by performing the threshold processing in which the portion below the appropriate threshold Th is set to zero. By performing threshold processing in this way, it is possible to prevent coding deterioration due to pixels with low significance,
More efficient and high-quality coded image processing can be performed.

In the present embodiment, the resolution of the image signal Si is converted based on the significant signal Sst subjected to the threshold processing, but the threshold processor 402 converts the threshold converted significant signal Sst into the encoder 405. May be directly output to and the threshold conversion significant signal Sst itself may be encoded.

(Fourth Embodiment) Next, referring to FIG. 14, an image signal coding apparatus E according to a fourth embodiment of the present invention.
C4 will be described. First, the image signal Si and the image signal S
The two-dimensional image signal Sv composed of the significant signal Ss corresponding to i is input. And a motion vector detector (M
E) 602 detects the motion vector and sends it to the motion compensator (MC) 604. In addition, an output from the motion vector detector (ME) 602,
Based on the output from a frame memory (FM) 611, which will be described later, the motion compensator (MC) 604 predicts the predicted image signal Sp.
Generate Predicted image signal Sp generated by the motion compensator
And a pixel value of the input image signal Sv, a difference value is calculated for each pixel by a difference device 603, and this difference value is orthogonal transform (DCT).
Quantizer (Q) 606 quantizes the component that has been transformed in 605 and orthogonally transformed. The quantized value is output to the variable-length encoder (VLC) 607 and inversely quantized by the inverse quantizer (IQ) 608 to obtain the inverse orthogonal transformer (IDC).
Inverse conversion is performed in T) 609.

The output of the inverse orthogonal transformer (IDCT) 609 thus generated is added by the adder 610 to the pixel value generated by the motion compensator (MC) 604, and the reproduced image signal Sv is obtained. As a threshold processor (Th) 61
2 is output. Further, the threshold value processor (Th) 612 determines only the reproduction significant signal Ss 'among the reproduction image signal Si' and the reproduction significant signal Ss 'included in the reproduction image signal Sv'.
Thresholding processing for converting a value equal to or less than the threshold value Sth into an insignificant value is performed to generate a significant signal Sst.

The significant signal Sst and the reproduced image signal Si '
Is a frame memory (FM) 611 as a decoded image signal.
Stored in. If the contents of the adder 610 are directly input to the frame memory 611, a significant signal that does not affect visually and a minute noise component will be used for the prediction of the next screen, and the efficiency of motion compensation will decrease, but Processor (T
h) By introducing 612, it is possible to prevent a decrease in efficiency. Furthermore, from the frame memory (FM) 611,
The decoded image signal is output to the motion vector detector (ME) 602 and the motion compensator (MC) 604.

The signal coded by the variable length coder (VLC) 607 is the coded image signal S of the image coding apparatus.
It is output as ic2.

With the above-described encoding device, the input signal Sv can be correctly encoded and the output signal Sic can be obtained without performing motion compensation on a significant signal that does not visually affect or a minute noise component. .

(Fifth Embodiment) Next, referring to FIG. 15, an image signal coding apparatus E according to a fifth embodiment of the present invention.
C5 will be described. The motion vector detector (ME) 602 detects the motion vector of the two-dimensional image signal Sv, and the motion vector is sent to the motion compensator (MC) 604. Further, the motion compensator (MC) 604 generates the predicted image signal Sp based on the outputs from the motion vector detector (ME) 602 and a frame memory (FM) 611 described later. It should be noted that the frame memory (FM) 611 receives the significant signal Ss ′ and the reproduced image signal S from the adder 610.
The reproduced image signal Sv 'composed of i'is directly input. The motion compensator (MC) 604 is a frame memory (F
M) 611 generates the predicted image signal Sp based on the significant signal Ss ′ and the reproduced image signal Si ′. Since the predicted image signal Sp generated by the motion compensator (MC) 604 includes a significant signal Ss ′ that has no visual effect, the threshold processor (Th) 701 reduces this significant signal Ss ′ to the threshold Sth or less. Is converted into an insignificant value to generate a threshold conversion predicted image signal Spt. In this way, significant signals that have no visual effect are removed, and coding efficiency is improved.

Output and input 2 of threshold processor (Th) 701
The pixel value of the three-dimensional image signal Sv is obtained by a differentiator 603 for each pixel, and the difference value is calculated by the orthogonal transformer (DCT) 6
05, the orthogonally transformed component is converted into a quantizer (Q).
Quantize at 606. The quantized value is a variable length coder (V
LC) 607 and the inverse quantizer (I
Q) Dequantize at 608, and inverse orthogonal transformer (IDCT) 6
Inverse conversion is performed at 09.

The output of the inverse orthogonal transformer (IDCT) 609 generated in this way is added by the adder 610 to the pixel value generated by the threshold value processor (Th) 701 to obtain a decoded image in the frame memory. It is stored in the (FM) 611.

The signal encoded by the variable length encoder (VLC) 607 is output as the encoded signal Sic of the image signal encoding device EC5. The encoding device EC5 as described above can efficiently encode the input image signal Sv signal by removing the significant signal that has no visual effect, and obtain the output encoded image signal Sic3.

(Sixth Embodiment) Next, referring to FIG. 16, an image decoding device DC according to a sixth embodiment of the present invention.
1 will be described. The image decoding device DC1 is used for decoding the coded image signal Sic1 generated by the image signal coding device EC3 shown in FIG.

The input signal encoded image signal Sic1 is input to the decoding device 802, and the significant signal Ss and the threshold signal S are input.
decrypted to th. The decoded significant signal Ss and threshold signal Sth are input to the threshold processor 805, threshold processing is performed, and the output thereof is used as the output signal Sv.

FIG. 17 shows an example in which the image quality is improved by the threshold value processing as C, D, E and F in FIG. C is the input significant signal Ss, and D is the transformed significant signal C. The output D is binarized by setting the value smaller than the threshold value Th to 0 and the other values to 1 and the value is E. The value smaller than the threshold value Th is rounded to 0, and the other values are left unchanged. What I did is F. When E is a binary significant signal and F is a multi-valued significant signal, it is possible to remove the significant signal that has no visual effect.

By performing the threshold processing as described above,
It is possible to remove a minute noise component from the decoded significant signal.

(Seventh Embodiment) Next, an image decoding apparatus DC2 according to a seventh embodiment of the present invention will be described with reference to FIG. The image decoding device DC2 is used for decoding the encoded image signal Sic3 generated by the image signal encoding device EC5 shown in FIG. The encoded image signal Sic3 is a variable length decoding device (V
LD) 902, where variable length decoding is performed.

On the other hand, the motion compensator (MC) 907 generates the predicted image signal Sp and sends it to the threshold value processor 906. The significant signal in the predicted image Sp generated by the motion compensator 907 is converted by the threshold value processor 906 into a value that is equal to or less than the threshold Th into an insignificant value to generate the converted predicted image signal Spt. The signal decoded by the variable length decoding device (VLD) 902 is input to the inverse quantizer (IQ) 903, is inversely quantized there, and is inversely orthogonally transformed by the inverse orthogonal transformer (IQ) 904. . The output of the inverse orthogonal transformer (IQ) 904 generated in this way is added by the adder 905 to the pixel value generated by the threshold value processor (Th) 906, and as a decoded image, the frame memory 908 (FM). ). As described above, the input encoded image signal Si
An output image signal Sv obtained by correctly decoding c3 can be obtained.

(Embodiment 8) Next, an image decoding apparatus DC3 according to an eighth embodiment of the present invention will be described with reference to FIG. The image decoding device DC3 is used for decoding the coded image signal Sic2 generated by the image signal coding device EC4 shown in FIG. The input signal Sic2 is a variable length decoding device (VLD).
It is input to 902, where variable length decoding is performed.

On the other hand, a variable length decoding device (VLD) 902
The signal decoded in (1) is input to an inverse quantizer (IQ) 903, is inversely quantized there, and is an inverse orthogonal transformer (IDC).
In step T) 904, inverse orthogonal transform is performed. Frame memory (F
The output of M) 908 is input to the motion compensator (MC) 907, and the predicted image signal Sp is generated. Motion compensator (MC)
The predicted image generated in 907 is added to the output of the inverse orthogonal transform (IDCT) 904 in the adder 905 and output as the image signal Sv, and at the same time, the threshold processor (T
h) 1001 converts a value equal to or less than the threshold value Th into an insignificant value and stores it as a decoded image Svt in the frame memory (FM) 908. Inverse orthogonal transformer (IQ) 9
The output signal of 04 is directly stored in the frame memory (FM) 90
When input to 8, the significant signal and the minute noise component that have no visual influence are output from the frame memory (FM) 908, and the accuracy of motion compensation by the motion compensator (MC) is reduced. The output of the frame memory (FM) 908 is input to the motion compensator (MC) 907, and the predicted image signal Sp is generated. The predictive image signal Sp generated by the motion compensator (MC) 907 is added by the adder 905 to the inverse orthogonal transform (IDC).
T) 904 is added to the output and output as an image signal Sv, and at the same time, is stored in a frame memory (FM) 908 as a decoded image Svt. As described above, it is possible to obtain the output image signal Sv in which the input signal 901 is properly decoded by preventing the motion compensation from being performed on the significant signal that does not visually affect or the minute noise component.

(Ninth Embodiment) Next, an image decoding apparatus DC3 according to a ninth embodiment of the present invention will be described with reference to FIG. Motion-compensated encoded input signal Sic
Is input to a variable length decoding device (VLD) 902, where it is variable length decoded. Variable Length Decoding Device (VLD)
The signal decoded in 902 is an inverse quantizer (IQ) 90.
3 and is inversely quantized there, and the inverse orthogonal transformer (I
DCT) 904 performs inverse orthogonal transform. The output of the inverse orthogonal transformer (IDCT) 904 thus generated is
The pixel value generated by the motion compensator (MC) 906 is added by the adder 905 and stored in the frame memory (FM) 907 as a decoded image. Since the significant signal of the decoded image output to the adder 905 includes a significant signal that has no visual effect, the threshold value processor (Th) 1101 converts the value equal to or lower than the threshold value Sth to an insignificant value.

As described above, the input coded signal Sci
It is possible to obtain the output image signal Sv that is correctly decoded. Further, in the case of FIG. 20, unlike the case of the threshold values in FIGS. 18 and 19, the threshold value St is irrelevant regardless of the encoding device.
By changing the value of h, it is possible to control a significant signal that has a delicate visual effect, and thus it is possible to change the image quality of an image to be displayed.

[0101]

As is apparent from the above description, by using the image conversion apparatus of the present invention, it is possible to perform image conversion by dividing significant pixels into insignificant pixels having no pixel value. The accuracy of the conversion process can be improved. Further, by using the image encoding device and the image decoding device of the present invention, it is possible to encode the pixel value of the insignificant pixel that does not affect the image quality of the reproduced image to a value that improves the encoding efficiency. , The encoding efficiency can be improved and its practical value is high.

In the present invention, resolution conversion processing such as pixel thinning or pixel interpolation that does not have the effect of insignificant pixels is performed on the input image signal to prevent deterioration of image quality due to the resolution conversion processing. In addition to this, high-quality and efficient encoding can be performed.

[Brief description of drawings]

FIG. 1 is a block diagram showing a structure of an image signal encoding device according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing an operation of the image signal encoding device shown in FIG.

FIG. 3 is an explanatory diagram regarding a method of determining a resolution conversion characteristic by the image signal encoding device according to the present invention.

FIG. 4 is an explanatory diagram showing an example of a function for the resolution conversion characteristic shown in FIG.

FIG. 5 is an explanatory diagram relating to image reduction by the image resolution conversion device according to the present invention.

FIG. 6 is an explanatory diagram related to significant signal complementation for color difference signals by the image resolution conversion device according to the present invention.

7 is a flowchart showing the operation of the image signal encoding device shown in FIG. 2 in more detail.

FIG. 8 is a flowchart showing a further operation of the image signal coding device showing the operation shown in FIG. 7.

FIG. 9 is an explanatory diagram relating to image enlargement by the image resolution conversion device according to the present invention.

FIG. 10 is an explanatory diagram related to resolution conversion of two images that are temporally consecutive according to the present invention.

FIG. 11 is a block diagram showing the structure of an image encoding device according to a second embodiment of the present invention.

FIG. 12 is a block diagram showing the structure of an image encoding device according to a third embodiment of the present invention.

FIG. 13 is a schematic diagram illustrating threshold conversion of a significant signal based on the present invention.

FIG. 14 is a block diagram showing the structure of an image encoding device according to a fourth embodiment of the present invention.

FIG. 15 is a block diagram showing the structure of an image encoding device according to a fifth embodiment of the present invention.

FIG. 16 is a block diagram showing the structure of an image decoding device according to a sixth embodiment of the present invention.

FIG. 17 is a schematic diagram showing an example in which the image quality is improved by the threshold value processing according to the present invention.

FIG. 18 is a block diagram showing the structure of an image decoding device according to a seventh embodiment of the present invention.

FIG. 19 is a block diagram showing the structure of an image decoding device according to a third embodiment of the present invention.

FIG. 20 is a block diagram showing the structure of an image decoding apparatus according to a fourth embodiment of the present invention.

[Explanation of symbols]

 Si image signal Ss significant signal Sc color signal Sr resolution conversion image signal SL resolution conversion characteristic selection signal Sth threshold signal Ti1 first input terminal Ti2 second input terminal To output terminal 103 resolution characteristic selector 105 resolution converter 106 output signal 301 resolution Transformer 402 Threshold processor 405 Encoder 406 Output signal 602 Motion vector detector 603 Difference device 604 Motion compensator 605 Orthogonal transformer 606 Quantizer 607 Variable length encoder 608 Inverse quantizer 609 Inverse orthogonal transformer 610 Adder 611 Frame memory 612 Threshold processor 701 Threshold processor 802 Decoder 805 Threshold processor 902 Variable length decoder 903 Inverse quantizer 904 Inverse orthogonal transformer 905 Adder 906 Threshold processor 907 Motion compensator 908 Frame Memory 10 1 thresholding unit 1101 thresholding unit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location H04N 11/04 9185-5C H04N 11/04 B

Claims (29)

[Claims] 1. A significant signal composed of a significant element (α) respectively representing a significant value (V) of the pixel (Sc) of an image signal (Si) composed of two or more adjacent pixels (Sc). An image signal encoding device (EC1) for encoding the image signal (Si) based on (Ss), wherein a significant value (α0) of a pixel (Sc; P1) in the vicinity of the pixel (Sc; P0). Resolution conversion characteristic determining means (103) for determining the resolution conversion characteristic (Pf; p) based on the above, and the image signal (S) with the determined resolution characteristic (Pf; p).
An image signal encoding device (EC), comprising: a resolution conversion means (105) for converting the resolution of i). 2. The image signal encoding device (EC) according to claim 1, wherein when the significant element (α) is binary, the resolution conversion characteristic (Pf; p) is expressed by the following equation. On the right side, p0 to pn are the respective pixel values of the image signal Si whose resolution is converted, and α is the referenced significant signal S.
An image signal encoding device (EC), wherein s has a significant element, and p on the left side can be represented by a pixel value of the image signal (Sir) converted by the expression on the right side. 3. The image signal encoding device (EC) according to claim 2, wherein the resolution conversion characteristic (Pf; p) is
The following expression α = α0 [+] α1 [+] ··· [+] αn−1, [+] shown on the right side can be represented by a logical sum, and an image signal encoding device (EC ). 4. The image signal encoding device (EC) according to claim 1, wherein when the significant element (α) is multi-valued, the resolution conversion characteristic (Pf; SL) is expressed by the following equation. P0 to pn shown on the right side can be represented by respective pixel values of the resolution-converted image signal Si, and w0 to wn pointers indicating whether or not the value of each significant element of the significant signal Ss referred to is 0. An image signal encoding device (EC) characterized by being capable of. 5. The image signal coding apparatus (EC) according to claim 2, wherein if at least one significant element (α) is significant, resolution conversion is performed with the resolution conversion characteristic. An image signal encoding device (EC). 6. The image signal encoding device (EC) according to claim 3, wherein if at least one significant element (α) is significant, resolution conversion is performed with the resolution conversion characteristic. An image signal encoding device (EC). 7. The image signal encoding device (EC) according to claim 4, wherein if at least one of the significant elements (α) is significant, resolution conversion is performed with the resolution conversion characteristic. An image signal encoding device (EC). 8. The image signal encoding device (EC: FIG. 5) according to claim 1, wherein the resolution conversion means (105).
Is an image signal encoding device (EC), wherein pixels of the image signal (Si) are thinned out. 9. The image signal encoding device (EC: FIG. 6) according to claim 1, wherein the resolution conversion means (105).
Is an image signal encoding device (EC), characterized by interpolating pixels of the image signal (Si). 10. The image signal encoding device (EC: FIG. 6) according to claim 1, wherein the resolution conversion means (105).
Is an image signal encoding device (EC), which produces a significant element (α) of the significant signal (Ss). 11. The image signal encoding device (EC2) according to claim 1, wherein the resolution conversion characteristic means (103).
Is provided upstream of the resolution conversion significant signal Ss to generate a resolution conversion significant signal (Ssr), and the resolution conversion significant signal (Ssr) is replaced with the resolution conversion characteristic means (103). The image signal encoding device (EC2) further comprising a second resolution converting means (301) for inputting to (1). 12. The image signal encoding device (EC3) according to claim 1, further comprising threshold processing means (402) for thresholding the significant signal (Ss) with a predetermined threshold (Th). Image signal encoding device (EC
3). 13. The image signal encoding device (EC3) according to claim 1, wherein N is used to perform an N-value conversion process on a pre-significant signal (Ss) corresponding to a parameter (Sth) input from the outside. Further, a quantizing means (402) and an encoding means (405) for encoding the image signal (Srb) whose resolution is converted based on the N-valued significant signal (Ss) and the parameter (Sth). An image coding apparatus (EC3) having the above. 14. The image signal encoding device (EC4) according to claim 1, wherein the motion prediction means (60) predicts a motion of the image signal (Si) and generates a prediction signal (Sp).
4), subtraction means (603) for subtracting the prediction signal (Sp) from the image signal (Si), and encoding means (605, 608) for encoding the difference signal of the subtraction means (603), Decoding means (60) for decoding the encoded difference signal
8, 609), and addition means (61) for adding the decoded difference signal and the prediction signal (Sp) to generate a decoded image signal (Sv ′).
0), and an N-valued means (612) for generating an N-valued decoded signal (Si ′ + Sst) for N-valued the decoded image signal (Sv ′).
And a memory (611) for temporarily storing the N-valued decoded signal (Si ′ + Sst) for use in the prediction signal generating means (MC). EC4). 15. The image signal encoding device (EC5) according to claim 1, wherein a prediction signal generation means (701) for generating a prediction signal (Sp) for the image signal (Si).
And the prediction signal (Sp) is N-valued to generate an N-valued prediction signal (S
N-valued conversion means (701) for generating pt), and the N-valued prediction signal (Spt) from the image signal (Si)
Subtraction means (603) for subtracting, the encoding means (605, 606) for encoding the difference signal of the subtraction means (603) to generate an encoded difference signal, and the decoding for decoding the encoded difference signal Means (608, 60)
9), and an addition means (6) for adding the output of the decoding means and the N-valued prediction signal (Spt) to generate a decoded image signal (Sv ′).
10) and the composite image signal (Sv ′) to the prediction signal generating means (6
04) for temporary storage in memory (61
An image coding apparatus (EC5) further comprising 1). 16. The N-value conversion means (40) according to claim 13.
2) is an image signal encoding device (EC3) characterized in that a value less than a predetermined value is set to 0. 17. The N-value conversion means (61) according to claim 14.
2) is an image signal encoding device (EC4) characterized in that a value less than a predetermined value is set to 0. 18. The N-value conversion means (70) according to claim 15.
1) is an image signal coding apparatus (EC5), wherein a value less than a predetermined value is set to 0. 19. The N-value conversion means (40) according to claim 13.
2) is an image signal encoding device (EC3) which is binarized using a predetermined value as a threshold. 20. The N-value conversion means (61) according to claim 14.
2) is an image signal encoding device (EC4) which is binarized using a predetermined value as a threshold. 21. The N-value conversion means (70) according to claim 15.
1) is an image signal encoding device (EC5) which is binarized using a predetermined value as a threshold. 22. An image signal (Sic1) encoded by the image encoding device (EC3) according to claim 13.
An image decoding device (DC1) for decoding the image signal, the decoding means (802) decoding the image signal (Ss) and the parameter (Sth) from the coded image signal (Sic1).
And an image signal decoding device (DC1) having N-value conversion means (805) for converting the decoded image signal (Ss) into an N-value corresponding to the parameter (Sth). . 23. An image signal (Sic) encoded by the image signal encoding device (EC5) according to claim 15.
3) An image signal decoding device (DC2) for decoding, the decoding means (902, 903, 904) for decoding the encoded image signal (Sic3), and the decoded image. Prediction signal (S
p) and a prediction signal generation means (907) for N-valued prediction signal (Sp) and N-valued prediction signal (Sp)
t), an N-value conversion unit (906), an addition unit (905) for adding the N-valued prediction signal (Spt) to the decoded image signal, and an output of the addition unit (905). The prediction signal generating means (9
07) for temporary storage in memory (90
8) An image signal decoding device (D)
C2). 24. An image signal (Sic) encoded by the image signal encoding device (EC4) according to claim 14.
An image signal decoding device (DC3) for decoding 2), comprising: decoding means (902, 903, 904) for decoding the encoded image signal (Sic2), and the decoded image. Prediction signal (S
p), a prediction signal generation unit (907), an addition unit (605) that adds the prediction signal (Sp) to the decoded image signal and outputs the result, and an output of the addition unit (905). It has an N-value conversion means (1001) for N-value conversion and output, and a memory (908) for temporarily storing the output of the N-value conversion means (1001) for use by the prediction signal generation means (907). An image signal decoding device (DC3) characterized by the above. 25. An image signal decoding device (DC4) for decoding a coded image signal (Sic) generated by coding a significant signal (Ss), the coded image signal (Sic) being decoded. Decoding means (902, 903, 904, 905) for converting, a prediction signal generating means (906) for predicting a motion of the decoded image signal, and the decoded image of the prediction signal (Sp) An adding means (905) for adding to the signal and outputting, an N-value converting means (1101) for converting the output of the adding means (905) into an N-value, and an output of the adding means (905) for the predicted signal generating means. (9
06) memory for temporary storage (90
7) An image decoding device (DC
4). 26. A significant signal composed of a significant element (α) respectively representing a significant value (V) of the pixel (Sc) of an image signal (Si) composed of two or more adjacent pixels (Sc). An image signal encoding method for encoding the image signal (Si) also based on (Ss), which is based on a significant value (α0) of a pixel (Sc; P1) near the pixel (Sc; P0). Resolution conversion characteristics (Pf; SL)
An image signal coding method, characterized in that a resolution conversion characteristic for determining is determined, and the image signal (Si) is resolution-converted with the determined resolution characteristic (Pf; SL). 27. A significant signal composed of a significant element (α) respectively representing a significant value (V) of the pixel (Sc) of an image signal (Si) composed of two or more adjacent pixels (Sc). An image signal encoding device (EC) that encodes the significant signal (Ss) based on (Ss) as well, and a significant value (α0 of a pixel (Sc; P1) near the pixel (Sc; P0). ), The resolution conversion characteristic determining means (103) for determining the resolution conversion characteristic (SL), and the significant signal (Ss) with the determined resolution characteristic (SL).
An image signal encoding device (EC1), comprising: 28. A significant signal composed of a significant element (α) respectively representing a significant value (V) of the pixel (Sc) of an image signal (Si) composed of two or more adjacent pixels (Sc). An image signal encoding method for encoding the significant signal (Ss) based also on (Ss), which is based on a significant value (α0) of a pixel (Sc; P1) near the pixel (Sc; P0). The resolution conversion characteristic for determining the resolution conversion characteristic (SL) is determined, and the significant signal (Ss) is determined by the determined resolution characteristic (SL).
An image signal encoding method, characterized in that the resolution conversion is performed. 29. The image signal encoding device (EC) according to claim 2, wherein the resolution conversion characteristic (Pf; p) is
The following expression α = α0 [×] α1 [×] ・ ・ ・ [×] αn-1, [×] shown on the right side is a logical product and can be represented by an image signal encoding device (EC ).