JPS5947912B2 - Color image encoding processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a color image encoding processing method, and particularly to a color image encoding processing method that reduces the amount of code per screen based on efficient encoding of a single image.

As a conventional color image encoding method, a predictive encoding method such as a DPCM method is mainly known for the purpose of highly efficient transmission of television signals.

In the DPCM method, a prediction signal is created using a sample value before the sample value to be encoded, and a prediction error signal, which is the difference between the sample value to be encoded and the prediction signal, is quantized and transmitted. Because the amplitude of the prediction error signal is concentrated in a narrow range due to the nature of the image signal, even if the number of quantization bits is reduced compared to normal PCM, the deterioration in image quality will not be so great. However, in these methods, it is necessary to send a prediction error signal for each sample point, and the number of bits required for quantizing the prediction error signal is limited by overload gradient noise, granular noise,
It cannot be reduced to an extreme level due to image quality deterioration factors such as false contours, and furthermore, since it is applied to general moving images, it is necessary to encode one screen within 33 ms, and the processing is limited to simple ones. There were some shortcomings.

The present invention is characterized by relaxing constraints on image encoding time and encoding color images by utilizing the two-dimensional local properties of images, and its purpose is to reduce the amount of code per screen. It is in.

FIG. 1 is an explanatory diagram for explaining the application concept of the encoding method according to the present invention, in which 100 is a target color image, P,, is a pixel in the i-th row and j-th column taken on the screen, and 101 -Y,l
Ol-I and lOl-Q correspond to the color image 100 when the color signal is separated into three components, Y, I, and Q. These are the image planes of each component corresponding to each component.

If the value of a pixel P, , is denoted by [P,,], an image is generally expressed by a two-dimensional distribution of [P,,].

Therefore, encoding an image means encoding the value [P,,] of each pixel. In addition to the above-mentioned YIQ separation method, a method of separating ζ into the three primary colors of red, green, and blue (R, G, and B) is also known as a method of expressing a color image. Here, general RG
Y, which is said to be able to reduce the amount of code compared to B separation,
Each image plane 101-Y, lOl after I and Q separation
-I, lOl-Q will be described as encoding. FIG. 2 is a diagram explaining the basic principle of the present invention,
200 in FIG. 2(a) corresponds to one plane (for example, 1 in FIG. 1) of a color image separated into Y, I, and Q.
01-Y), and Bk is one plane 2
201 in FIG. 2(b) is an enlarged view of the extracted block Bk.

In the encoding method according to the present invention, each plane of Y, I, and Q is encoded in units of blocks 201. That is, one plane is encoded by dividing it into a large number of blocks and including, for example, 16 pixels P, -P, f3, f, which are included in one block Bk. The signal value of [
P,,]~[P,,f. ,,f. ] are the following tone components AkO, akl and resolution component elements, , ~R, l3,,
Encode using f3. Bk here.

is a subset of Bk consisting of pixels P, , whose signal value [P, ,] is smaller than the threshold Tk among the pixels in Bk, and Bk is a subset of Bk consisting of the pixels P, , whose signal value [P, ,] is smaller than the threshold Tk.
Mk. , mk, are each a subset Bk. , Bk, represents the number of pixels included in ,Bk. At this time Ak. , ak, are called tone components of block Bk. In this case, the pixels of Bk are two mutually exclusive subsets Bk. , Bk, can be classified into two types. Therefore, Bk of pixels within Bk. When assigning the code ゜0'' to the pixels belonging to Bk, and the code ``1'' to the pixels belonging to Bk, for each pixel P, use this code to R, ,~R,f3,,f,
The above codes ``0'' and ``l'' are respectively given as follows. This is called the resolution component. The basis of this encoding is to approximate the brightness of any pixel in block Bk with either the AkO or akl signal obtained by equation (1) in this way, and the encoded output is basically Gradation component Ak.

, ak, and the resolution component element. ,~R,f3,,f3(
(If one block has 4×4 pixels) is given. Also,
Of these, the gradation components AkO and akl are given only from the luminance of pixels in block Bk, but these values are considered to have a strong correlation with the gradation components of adjacent blocks Bk-1 and B11. Therefore, the tone component Ak. ,ak,
Instead, the sum component Akm, the difference component Akd, and the difference Dk. Let us define n,
The state of block Bk can be expressed by Akd and dkm.

In this way, Akd and Dk. n is AkO, akl
Compared to , the amplitude distribution is concentrated near O, and AkO, akl
It is possible to encode with a smaller number of bits than to encode it as is. By performing the above-described processing on each component of Y-1-Q, resolution components of tone components are obtained as encoded outputs. The present invention is based on the principle described above, and further includes means for varying the size of the block in accordance with the local characteristics of the image.

In other words, in areas with large brightness changes in the image, the deterioration caused by encoding is suppressed by reducing the block size.
Furthermore, in areas where brightness changes are small, the amount of code is reduced by performing encoding with a larger block. Here, using Figure 3, we will explain how to make the block size variable.
The principle of sequentially dividing a block, which is a method, will be explained. FIG. 3 shows the principle for realizing the present invention. In the figure, 301 is a block having 8×8 pixels;
02 to 305 are blocks with 4 x 4 pixels, 30
6 to 309 indicate blocks having 2×2 pixels;
An example of obtaining three types of blocks is shown below.

First, the Y component plane of the original image is divided into blocks each having 8×8 pixels. Focusing on one of the blocks 301, this block is subjected to the aforementioned encoding process, that is, the encoding process of separating and encoding the gradation component and the resolution component. Encoding process. The difference ε,j between the Y component of the pixel after conversion and the Y component of the pixel before conversion is calculated for all pixels in the block 301. Next, the sum ε21 of ε and 3 squared ε20 for all pixels in the block is determined. In other words.

Next, a preset threshold value ε21 for varying the block size is compared with ε21, and the dividing variable D1 is determined.
Establish. When D1=0, the conversion error ε21 due to the encoding process is smaller than the threshold value ε210 for varying the block size, so the block 301 is encoded as is.

However, when D1=1, it is assumed that the conversion error ε21 in the pixel block 301 is too large, and as shown in FIG.
Divide into 4 parts into 305 parts. Each of these four divided blocks 30
2 to 305 as well, the same operation as above is performed to obtain the conversion errors, and the total error ε22 is obtained.

Similarly, the division variable D2 is determined by comparison with a predetermined threshold value ε22T for varying the program. That is, .

When D2=O, encode with the same block and D2
When = 1, the division is repeated again, for example, the block is divided into four blocks each having 2×2 pixels, and the encoding process is performed on each of the blocks 306 to 309 divided into four in this way. get the output. As mentioned above, by dividing the block, encoding can be performed with the conversion error below a certain value, and the variable D for the timetable can be encoded.
If l and D2 are stored as information for each block, the information of the Y component plane can be easily restored. The reason for performing the above operation for the Y component is that, as is well known as the property of continuous television signals, the three components Y, I,
This is because the Y component has the largest amplitude change in Q and has the strongest effect on resolution, and the Y component and I, Q
The size of the block used for the component can be set equal to or larger than that of the I and Q components compared to the block of the Y component, and the block of the I and Q components can be set larger than the block of the Y component. This enables even more efficient transmission.

I can choose a larger block size.
This is because the degree of luminance change is generally not as large as that of the Y component, just as Q does not pose much of a problem even if it is band limited in normal television transmission. As described above, in the present invention, for each of the Y-1-Q components, the aforementioned gradation component, resolution component, and Y component, the variables Dl and D2 for division are transmitted as encoded outputs or converted into files. I will do it. FIG. 4 is a block diagram of an encoding processing device using the present invention, in which 400 is a target color image;
401 is an image input section; 402 is a Y, I, Q separation section; 40
3Y, 403, and 403Q are A for sampling the Y, I, and Q separated signals and converting them into digital signals.
/D converter, 404Y, 404, 404Q are buffer memories that can store pixel information for at least l blocks of each signal, 405Y, 405, 405Q are code codes for each signal of Y, I, Q, respectively. 406 is an encoding control unit for performing the above-mentioned block division control, and 407 is a code storage file unit.

Next, the operation of the present encoding processing device will be explained using FIG. First, a color image 400 is scanned by an image input section 401 and sent to a YIQ separation section 402 . In this case, an ordinary color TV camera or the like can be used as the image input unit 401. In addition, when the three primary color signals R, G, and B are input, the YIQ separator 402 uses normal RGB-
This is for YIQ exchange. Next, the signals separated into YIQ components are sent to A/D converters 403Y and 40 for each plane.
3, 403Q, it is sampled and converted into a digital signal, and then sent to the buffer memory 404Y, 404, 404Y.
It is written to 04Q as pixel brightness information.

The role of the buffer memories 404Y to 404Q is to store pixel luminance information for each block in order to match the next encoding process when the image input unit is of a raster scanning type like a normal television camera. It has a buffer capacity for multiple scanning lines corresponding to the size of the block. For example, as mentioned above, the block sizes are Y, I.
, Q If each plane has a maximum of 8×8 pixels, it is sufficient to have memory for 8 scanning lines. The pixel luminance information stored in buffer memories 404Y to 404Q is sent to encoding units 405Y to 405.

and is sent to the encoding control unit 406 in units of blocks, and according to the aforementioned encoding principle of making the block size variable,
Converted into gradation components and resolution components. At this time, the encoding unit 4
05Y~405.

is the division variable Dl output from the encoding control unit 40},
D2 processes blocks of 8 x 8 pixels, 4 x 4 pixels, 2 x 2 pixels, etc., but these processing operations basically change the size of the blocks by simply performing the operation of equation (1). Regardless of the size, it can be easily realized using existing circuit technology. The encoding control unit 406 also detects the error in equation (3),
The main function is to obtain the division variables D, D2 shown in equations (4) and (5), and the existing multipliers, adders,
This can be easily realized using a comparator or the like. The information encoded in this manner is stored in the code storage file section 407 together with the division variables Dl and D2.

In the configuration of this encoding device (FIG. 4), buffer memories 404Y to 404Q) encoding units 405Y to 4
05Q is shown in a form where each component of Y, I, and Q is configured independently, but depending on the memory capacity and processing capacity of each part, it is not necessarily necessary to prepare them independently, and only one processing It is also possible to use the system in a time-division manner, and it is also obvious that the above processing functions can be realized by software using a microprocessor or the like. FIG. 5 shows a block diagram illustrating the configuration of the decoding apparatus described above to decode the codes stored in the code storage file 407 and obtain a display image signal.

In the figure, 501Y to 501Q are buffer memories corresponding to planes of Y, I, and Q components, 502Y to 502Q are decoding units, 503 is a decoding control unit, 504 is a color signal synthesis unit,
505 is a color image display section. The operation of the decoding processing device will be explained using FIG.

The information stored in the code storage file 407 is temporarily stored in each of the buffer memories 501Y to 501Q for each plane. At this time, the buffer memory capacity is at least 1
It is only necessary to be able to store information for a block. Also Y
The component buffer memory 501Y also stores division variables D, D2 obtained during encoding. Variables Dl and D2 for this division are sent to the decoding control section 503, and signals corresponding to the actual gradation components and resolution components are sent to the decoding sections 502Y to 502Q. Decoding control unit 50
3, the block size is detected and this information is sent to the decoder 50.
2Y~502.

The decoding unit 502 extracts gradation components and resolution components according to the size of the block, and decodes the luminance information of each pixel within the block. Information on each pixel in the block decoded for each plane is input to a color signal synthesis section 504. The color signal synthesis section includes a buffer memory, and temporarily stores the input luminance information in the buffer memory. Next, the brightness information in the buffer memory is sequentially read out in accordance with the scanning speed and scanning format of the image display unit 505, and the image signals are combined and restored by the image display unit 505 to obtain a desired image. In this decoding processing device as well, although a format in which the decoding section is configured independently for each plane has been described, it is also possible to use one processing system in a time-sharing manner.

The present invention has been described above, and in the main text, a method of sequentially dividing the block was explained as a method of varying the block, but a method of expanding the block is also considered, and the shape of the block can also be changed to a square. Not only rectangular blocks, but also rectangular blocks are acceptable.

In addition, although the conversion error ε2 is used as a parameter for making the program size variable, it is not limited to this. For example, it may be the sum of the absolute values of errors caused by conversion, or the program size can be changed depending on the magnitude of the luminance change. Considering switching the size of the block, the size of the block may be changed based on the difference value between the tone components AkO and akl, for example, the value of Akd described in equation (2). Furthermore, in areas with little brightness change, 8
Instead of encoding using a large block such as a x8 pixel block, it is also possible to use four 4 x 4 pixel blocks and each block may be displayed using only one luminance signal.
In this case, the resolution component is also unnecessary, and Ak. It is possible to use only n as luminance information. In the above description, it has been shown that the encoded signal is filed, but it goes without saying that the present invention can be applied to a transmission system that directly transmits the encoded signal via a transmission path.

As explained above, according to the present invention (according to which the size of the block is varied according to the local characteristics of the image), it is possible to obtain high efficiency and high quality image as an encoding method. When used in an image file system such as the one described above, the amount of files can be reduced, and when used in a communication system, it has the advantage of shortening transmission time.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram for explaining the application concept of the encoding method according to the present invention, FIG. 2 is an explanatory diagram for explaining the basic principle of the encoding method according to the present invention, and FIG. 3 is an explanatory diagram for explaining the encoding method according to the present invention. 4 is a block diagram showing the configuration of an encoding processing device using the encoding processing method of the present invention, and FIG. 5 is a diagram showing the decoding processing using the encoding processing method of the present invention. FIG. 2 is a block diagram showing the configuration of the device. 100...Target color image, 101Y~
101,... Image plane of each component for color image 100 when the color signal is separated into three components Y, I, Q, 200... One of the image planes of each component , 201...Procs that are image processing units, 301...Procs having 8x8 pixels, 302-305...Procs having 4x4 pixels, 306-- 309...Prok having 2×2 pixels, 400...Target color image, 401...Image input section, 402...
YIQ separation section, 403,~403,...A/D
Conversion section, 404Y to 404c...Buffer memo I
8405, to 405Q... Encoding section, 406.
...Encoding control section, 407...Storage file section, 501Y to 501, ...Buffer memory, 502Y to 502, ...Decoding section, 503.
...Decoding control unit, 504...Color signal synthesis unit, 505...Image display unit.

Claims (1)

[Claims] 1 Y that decomposes the color image signal into three components: Y, I, and Q
an IQ separation section, an A/D conversion section that samples signals for each component and converts them into digital signals, and a buffer that temporarily stores digital image signals corresponding to each component obtained by the A/D conversion section. The image signals of each component having a two-dimensional distribution stored in the memory section and buffer memory are
Divide into dimensional small blocks, calculate the average brightness value of the brightness of each pixel belonging to each block for each block, use the average brightness value as a threshold, and set a brightness value greater than the threshold value for each pixel in the block. A set of pixels having a luminance value smaller than the threshold value is divided into a set of pixels having a luminance value smaller than the threshold value, and a gradation component corresponding to the average luminance value of each of the two sets is determined, and the luminance value of each pixel included in the block is calculated. means for replacing the luminance value of each pixel with the value of the gradation component to obtain a converted luminance value based on the magnitude relationship between the pixel and the threshold; and a resolution component corresponding to the information indicating which of the two sets each pixel belongs to. means for obtaining the sum and difference values of the gradation components, furthermore, the sum of the values obtained by the same means as in the case of obtaining the gradation components for blocks adjacent to the block, and the above-mentioned gradation of the block. means for calculating the difference between the sum of the tone components, and encodes the difference information between the small blocks of the value of the sum of the tone components, the difference information of the tone components, and the resolution component. The difference between the original luminance value of the block and the transformed luminance value obtained by the encoding processing unit is calculated for the encoding unit and the Y component to be outputted, and the sum of the squared error values within the block is calculated in advance. The total sum is compared with a predetermined threshold value for making the block size variable, and if the sum is larger than the threshold value for making the block size variable, the block size is made smaller; A control signal for performing encoding processing on each of the Y, I, and Q components with the size of the block unchanged or increased is output to the encoding section for each of the Y, I, and Q components; 1. A color image encoding processing method, comprising: an encoding control unit having a function of variably controlling a block size and a function of outputting the control signal together with the encoded output.