JPH02174489A - Color picture coding system - Google Patents

Abstract

PURPOSE:To code a color picture efficiently by coding luminance information so as to keep the resolution of a range of a picture reproduction capability range of a recording system and coding chrominance information with a resolution in response to the color discriminaction threshold level. CONSTITUTION:A 3 primary color signal is fed sequentially from a picture input device 1, the converter 2 converts the signal into a luminance signal L* and chrominance signals a*, b* based on uniform space and stores them once. The signal L* eliminates a smaller component in the AC component of a matrix subjected to cosine conversion, decreases a total quantization bit number and reduce the information quantity. On the other hand, the signals a*, b* eliminate a high frequency component from each picture element matrix and eliminates the information unsensible because the level is smaller than the color discrimination threshold level. A line buffer 3 sends the luminance information from the converter 2 to a 1st encoder 4 and sends the color information to a 2nd encoder 5. The encoder 4 applies Huffman coding and the encoder 5 uses a lookup table to encode the signal and the result is fed to a buffer and selector 6.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention relates to a color image encoding method for efficiently encoding color images.

(Prior Art) A color image has a large amount of information compared to an image with only shading such as black and white, so if the image information of a color image is transmitted as it is, it will take a lot of time and cost due to the use of a line.

Therefore, there is a need to efficiently encode and compress image information of color images.

Generally, color images are R (red), G (green),
It is represented by signals representing the three primary colors of B (blue), but if these signals were to be encoded and compressed all at once, the scale of the hardware would simply become excessive.

Also, the three primary color signals may be encoded separately, but although the signal characteristics of R, G, and B are slightly different from each other, there is almost no difference in characteristics such as spatial frequency, so it is not effective. It is difficult to perform specific compression. Therefore, the three primary color signals are linearly converted to form a ycbc signal, a Y signal indicating 11 degrees, and a Cb, Cr signal indicating the color difference.
Encode each signal separately.

In this case, since the spatial frequencies of the Cb and Cr signals are relatively low, effective compression is possible.

However, encoding is not done in consideration of human visual characteristics, such as the linearity of luminance shown by the Y signal with respect to visibility, and the color discrimination threshold for color differences shown by the Cb and Cr signals, and the recording system etc. No consideration was given to the image reproduction ability of the system. As a result, visually effective data may be lost in the reproduced image, or unnecessary data may remain.

(Problem to be Solved by the Invention) Conventionally, when encoding color images, human visual characteristics and the image reproduction ability of the recording system were not taken into consideration. There was a problem that it did not work.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a color encoding method that can efficiently encode color images by taking into account human visual characteristics and the image reproduction ability of a recording system.

[Composition of light]

(Means for Solving the Problems) The present invention provides a conversion means for converting a color image represented by three primary color signals into luminance information and chromaticity information based on a uniform color space, and converting the chromaticity information into a resolution within an image reproduction capability range. The present invention is characterized in that it includes a first encoding means that encodes the chromaticity information so as to preserve it, and a second encoding means that encodes the chromaticity information at a resolution that corresponds to a color discrimination threshold.

(Function) According to the present invention! ! The information is encoded so as to maintain the resolution within the image reproduction capability range of the recording system, and the chromaticity information is encoded at the wI quadrature that corresponds to the color discrimination threshold, reducing the amount of information more than necessary. A color image can be efficiently encoded without performing the above process or producing an excessive amount of information below the color discrimination threshold.

(Embodiments) Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram showing an encoding device to which an embodiment of the color image encoding method according to the present invention is applied. In the figure, the image input device 1 performs decoding using, for example, a COD (charge-coupled device) image sensor, and sequentially forms three primary color signals corresponding to each pixel of one line in each main scan. Then, these three primary color signals are sequentially applied to the converter 2.

These three primary color signals are the three primary colors of the pixel: R (red), R
This is a digital signal indicating the brightness of (green) and B (blue).

The converter 2 sequentially inputs the three primary color signals, converts the three primary color signals into signals L*a*b* based on the uniform color space of CIE1976L*a*b*, and temporarily stores the signals. here,
The signal * indicates the degree of pixel of the color image, and each signal a*b* indicates the chromaticity of the pixel.

Next, the converter 2 accumulates each of the signals L*a*b* for, for example, n lines, and then cosine transforms the pixel matrix L*ij of n rows and n columns for luminance information shown in FIG. In other words, the pixel matrix L*ij represents the luminance indicated by the signal * for each pixel, and such a pixel matrix L
When *ij is cosine transformed *layer, matrix L IJ is formed.

Here, cosine transformation is generally expressed by the following equation (1).

F=CXC' However, X is a pixel matrix, C is a cosine transformation matrix, C1
is the transposed matrix of matrix C, and F is the transformed matrix.

The transformed matrix F has the following characteristics.

(1) The components in the first row and first column of the matrix F indicate the sum of DC components included in the pixel block corresponding to the pixel matrix X.

(2) Each component in matrix F is uncorrelated.

(3) AC components (components other than the first row and first column) in matrix F are not large.

Therefore, if small components are removed from the alternating current components in the cosine-transformed row*layer/column L IJ shown in Figure 2, the total number of quantization pits can be reduced without significantly affecting the image. The amount of information can be reduced by

Therefore, the converter 2 deletes each component related to the threshold line A and each component on the lower right side of the threshold line A as small AC components in the cosine-transformed matrix LIJ shown in FIG. This threshold line A is set in advance so that information out of the current capability of the recording system is deleted from the matrix L+j when the image is reproduced.

Such cosine layer transformation of the pixel matrix L*ij of luminance information and deletion of small alternating current components from the cosine-transformed batting average L+j are performed for every n row and n column for n lines. Data bit signals for one line of a color image indicating brightness information are sent to the line buffer 3.
will be added sequentially.

On the other hand, the converter 2 removes high frequency components from each pixel matrix a*ij, b*ij of n rows and n columns regarding the chromaticity information shown in FIG.
Form each pixel matrix a + J, b IJ in the layer. Here, the pixel matrix a*+j contains each signal a for each pixel.
It represents the chromaticity indicated by *, and the image matrix b
IJ represents the chromaticity indicated by *... for each signal b* for each pixel. When high frequency components are removed from each of these pixel matrices a*ij, b*ij, fine color changes in the color image are eliminated, and information that is smaller than the color discrimination threshold and cannot be detected is deleted.

Each pixel matrix a''ij of such chromaticity information.

By removing high frequency components from b*ij, each pixel matrix a”
ij, b'' ij is formed every n rows and n columns for n lines, and the data bit signals for the color image Qn lines indicating the formed chromaticity information are sequentially added to the line buffer 3. .

The line buffer 3 is constructed using a FIFO circuit, and receives the I! from the converter 2! When 0,592 data bit signals indicating ili degree information are sequentially input, these data bit signals are sent to the first encoder 4. Also, converter 2
When 0592 data bit signals indicating chromaticity information are inputted sequentially, these data bit signals are sent to the second encoder 5.

When the first encoder 4 sequentially inputs the data bit signals corresponding to the row*layer/column L IJ shown in FIG. Huffman coding is performed based on the data, and a digital signal CD1 indicating the luminance information of the Huffman code thus formed is applied to the sofa/selector 6.

Also, the second encoder 5 has *layer...
*Layer... When data bit signals corresponding to each pixel matrix a + J, b IJ shown are input in sequence, these data bit signals are encoded. This encoding is performed using a lookup table built into the second encoder 5, and this table shows color discrimination thresholds as the chromaticity increases on the coordinates associated with each chromaticity a, b. The encoded data corresponding to the property that the ellipse expands is stored in advance. Add to this J2, each pixel matrix a" ij, b"
ij is encoded in such a way that information that is smaller than the color discrimination threshold and cannot be sensed is deleted regardless of the chromaticity. The digital signal CD2 representing the chromaticity information thus encoded is applied to the buffer/selector 6.

When the buffer/selector 6 receives the digital signal CD1 and the digital signal CD2, it outputs these digital signals C.
D1 and the digital signal CD2 are multiplexed and sent to the transmission line 7.

In this embodiment, the three primary color signals are converted into each signal L*a*b* based on a uniform color space, and the components that are out of the image reproduction capability of the recording system are removed from the luminance information signal L* for encoding. At the same time, information that is smaller than the color discrimination threshold and cannot be detected is deleted from each signal a and b of the chromaticity information * * before encoding. For this reason, color images can be encoded efficiently, and even if the recording system decodes and reproduces the image, visually effective data may be lost in the image.
No unnecessary data remains.

Note that in this embodiment, the pixel matrix L*i for luminance information
j is subjected to cosine transformation, but other than cosine transformation, Hadamard transformation, Barr transformation, etc. can be exemplified.

〔Effect of the invention〕

As explained above, according to the present invention, luminance information is encoded so as to maintain the resolution Q degree of the image reproduction capability range of the recording system, and chromaticity information is encoded with a resolution degree according to the color discrimination threshold. , so the information Q is reduced more than necessary,
Also, it does not generate excessive information below the color discrimination threshold.
Thus, it is possible to provide a color image encoding method that can efficiently encode color images.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an encoding device to which an embodiment of the color image encoding method according to the present invention is applied, and FIG. 2 is a diagram used to explain encoding of luminance information.
FIG. 3 is a diagram used to explain the encoding of chromaticity information. DESCRIPTION OF SYMBOLS 1... Image input device, 2... Converter, 3... Line buffer door, 4... First encoder, 5... Second encoder, 6... Buffer 1 also Selector, 7... transmission line.

Claims (1)

[Scope of Claims] Conversion means for converting a color image represented by three primary color signals into luminance information and chromaticity information based on a uniform color space, and encoding the luminance information so as to preserve resolution within an image reproduction capability range. A color image encoding method comprising: a first encoding means for encoding the chromaticity information; and a second encoding means for encoding the chromaticity information at a resolution according to a color discrimination threshold.