JP2930290B2 - Color image processing equipment - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color image processing apparatus for efficiently compressing color image information. 2. Description of the Related Art Conventionally, when encoding a luminance signal and a chromaticity signal of a color image, if encoding is performed for each pixel, the amount of encoded data becomes very large. The amount of encoded data has been reduced by dividing the image into blocks and sharing the encoded data to some extent in the blocks. [0003] However, in the above-mentioned prior art, when encoding a chromaticity signal in block units,
If encoding is not successful with the initially set block unit size, for example, if encoding with the initially set block size causes significant image degradation to the human eye, the first set block is Must be changed to an encoding target area having a different size. SUMMARY OF THE INVENTION The present invention has been made in view of the above conventional example, and is directed to a color image processing apparatus for encoding by switching the size of an encoding area in a chromaticity signal of a color image signal to an optimal size. It is an object to perform efficient coding in consideration of the above. It is a further object of the present invention to provide a color image processing apparatus capable of generating a code amount corresponding to the size of a coding area. [0005] In order to achieve the above object, a color image processing apparatus according to the present invention has the following arrangement. That is, generating means for generating a luminance signal and a chromaticity signal from a color image signal (the color separation circuit 1 in FIG.
And equivalent) to 07, the chromaticity signal, the predetermined area (first encoding mode (in the embodiment averaging circuit for each coded is divided into equivalent) in blocks of 4 × 4 pixels in Example 110 , LUT 112) and the predetermined area determined based on the content of the luminance signal (corresponding to the luminance signal L * input to the average calculation unit 121 and the area determination unit 120 in the embodiment). A second encoding mode (in this embodiment, an average value) in which the image is further divided into a plurality of partial regions smaller than the region (corresponding to each region in which 4 × 4 pixels are described in two sentences by the region determination unit 120). Circuit 11
1, corresponds to the case of using the LUT 112) and the selector 109 in the coding means for operating selectively (Example,
(Equivalent to averaging circuits 110 and 111 and LUT 112)
And the code amount when the predetermined region is coded in the first coding mode (corresponding to the code amount of 8 bits of C0 in FIG. 3 in the example) in the second coding mode. It is characterized by having control means for controlling so that each code amount (corresponding to 4 bits of C1 or C2 in FIG. 3 in FIG. 3 in the embodiment) becomes smaller when each is encoded. Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. [Description of Image Processing Apparatus (FIGS. 1 and 2)] FIG. 1 is a block diagram of an image processing apparatus according to the present embodiment. In FIG. 1, reference numeral 101 denotes a glass document table 102.
This is a color original placed downward on the top. The color original 101 is illuminated from below by the lamp 100 and is imaged on the R, G, B color separation sensors 104 by the rod lens array 103. In this manner, the sensor 104 reads the color document 101 by photoelectric scanning, and reads the three primary colors (RG
B) is converted into an analog image signal and output. Sensor 1
Each of the R, G, and B analog image signals from S.04 is held for a predetermined time by a sample-and-hold circuit (S / H) 105. Based on the held voltage value, an A / D converter 106 converts the analog image signals into eight for each color. It is converted to bit digital data. Reference numeral 107 denotes a color separation circuit that separates each input 8-bit signal into a luminance signal and a color difference signal.
It is decomposed into three-dimensional three-dimensional coordinates in a uniform color space of CIE 1976, and brightness L * is used as a luminance signal, and chromaticity coordinates a * and b are used as a color difference signal.
* Is what you get. Note that this color separation circuit is not limited to color separation into the L * a * b * color space.
Other color sequences such as Q and L * u * v * of CIE 1976 may be used. Reference numeral 108 denotes chromaticity coordinates a * and b * from the color separation circuit 107, which detects the magnitude of a change in chromaticity within a block composed of a plurality of pixels set in advance, and detects the color of the image data. This is an edge detection unit that detects an edge based on. The detection of this chromaticity change is performed, for example, by a *, b *
Is subjected to a differentiation process or the like in the block, and the magnitude of the change is detected by comparing the differential value with a predetermined threshold value. Hereinafter, a block having a chromaticity change larger than the threshold is referred to as an edge block, and a block smaller than the threshold is referred to as a flat block. A switch 109 receives the detection result 118 from the edge detector 108, selects the a * and b * signals to either the averaging circuit 110 or 111, and inputs the selected signal. ) Is input as a * and b * data 119 to the averaging circuit 110 when the block is determined as a flat block.
It operates so that 9 is input to the averaging circuit 111. The averaging circuit 110 calculates an average value of each of a * and b * for all the pixels in the block, and
Numeral 11 divides the block into two regions, and calculates the average value of a * and b * for each region. A method of dividing into two areas is, for example, a lightness signal L * and chromaticity signals a *, b *.
By focusing attention on the correlation at the edge portion. Normally, L * and a *, b * have low correlation, and therefore, even if the spatial resolution of a *, b * is reduced, the image is not much degraded. The distribution of L * and a *, b as shown in FIG.
It can be considered that the distribution of * is almost the same. FIG. 2A is a graph showing the distribution of L *, and FIG. 2B is a graph showing the distribution of a * in shades of 4 × 4 pixel blocks (the distribution of b * is almost the same). is there. FIG. 2 (A),
In (B), one cell 21 corresponds to one pixel,
A block is composed of 4 × 4 pixels. The operation of the averaging circuit 111 will be described focusing on the above points. First, an average value L * / of L * in a block of interest is obtained by an average calculation section 121, and a pixel block of interest is obtained by an area determination section 120. The magnitude of the L * value of the pixel within the pixel is compared with the average value L * /, and the block of interest is set to the region of L * ≧ L * /
* <L * / area. The averaging circuit 111 calculates an average value of a * and b * in an area of L * ≧ L * / and an average value of a * and b * in an area of L * <L * / based on the area information. calculate. In this way, the average value of a * and b * of the block of interest is output from the averaging circuit 110 or 111 and is input to the look-up table (LUT) 112 which is an encoding table. Hereinafter, the processing after the LUT 112 will be specifically described. The chromaticity signal input to the LUT 112 is an average value (a0 * /, b0 *) of a set of a * and b * of a flat block.
/), And two sets of a *,
The average value of b * (a1 * /, b1 * / and a2 * /, b2 * /) is obtained. Therefore, in the LUT 112, each of the two sets of edge blocks a1 * /, b1 * /, a2 * /, b2 * /
By encoding with a bit number that is half of a0 * / and b0 * / of the flat block, the information amount of each flat block and edge block can be made equal.
That is, for example, in a flat block, (a0 * /, b0 * /) is regarded as one vector and vector-quantized to an N-bit code value C0, and in an edge block, (a1 * /, b1 *)
/) Is regarded as one vector, and vector quantization may be performed on an N / 2-bit code value C1 and (a2 * /, b2 * /) may be vector-quantized on an N / 2-bit code value C2. FIG. 3 is a diagram showing an example of coding of a flat block and an edge block when N = 8. Reference numeral 31 denotes an 8-bit code (C0) obtained by coding a flat block, 32 denotes an 8-bit code obtained by coding an edge block, and C1 and C2 denote 4-bit codes corresponding to respective areas. I have. Latches 115 and 116 shown in FIG. 1 are latches for creating such a code. Latch 115 is a 4-bit latch circuit, and 116 is an 8-bit latch circuit. Reference numeral 117 denotes a compression circuit for performing a compression process on the luminance signal L * by a known method such as orthogonal transformation or vector quantization. [Description of block vector quantization (encoding) (FIGS.
FIG. 6) shows the LUT 112 and the latch circuits 115 and 11
FIG. 6 is a diagram showing a connection with the sixth embodiment. The LUT 112 is an averaging circuit 110 or 1
11 and the average value 114 of a * and b * from the edge detection unit 10
8 and a 1-bit detection signal 113 are input.
The block of interest is a flat block (detection signal 113 =
0), the LUT 112 stores an 8-bit code (C
0) for two clock cycles. Code (C0)
Are latched by the latch circuit 115 in the first clock cycle, and the 8-bit code (C0) is latched in the latch circuit 116 in the second clock cycle. When the target block is an edge block (detection signal 113 = 1), the LUT 112 sets both the upper 4 bits and the lower 4 bits to C in the first clock cycle.
A code of 1 is output, and a code in which both the upper 4 bits and the lower 4 bits are C2 is output in the second clock cycle. As a result, the code composed of C1 (4 bits) and C2 (4 bits) indicated by 32 in FIG. 3 is latched in the latch circuit 116 in the second clock cycle. FIG. 5 shows a *,
FIG. 4 is a diagram illustrating a distribution of a b * chromaticity signal. The vertices 401 to 406 of the hexagon are the six subtractive primary colors, respectively, and are yellow (Y), green (G), cyan (C), blue (B), magenta (M), and red (R). It corresponds to the degree coordinate. Therefore, in order to quantize a * and b * of the flat block, the inside of the hexagon in FIG.
(8-bit) area, and a look-up table for assigning an 8-bit code to each area is provided in the LUT 11.
2 is provided. On the other hand, in the edge block, the hexagonal area shown in FIG. 5 is divided into 16 (4 bits) areas, and a look-up table L is assigned so that a 4-bit code is assigned to each area.
The UT 112 must be configured. Therefore, region division is performed focusing on the following points.色 At the edge of the image, the color reproducibility for the input document is not very important. ⇒ In the edge part of the image, 6 of Y, M, C, R, G, B
It is often composed of primary colors and any of white and black. (4) At an edge portion with low saturation, if the area division is made coarse, a change in color is more conspicuous. An example of area division focusing on the above points is shown in FIG.
Shown in In this figure, the area is divided into 16 areas 501 to 516, but it is finely divided near the achromatic color (513).
516) It can be seen that the division in the direction connecting the origin and the apex of the hexagon is finely divided. By storing the look-up table having the above characteristics in the LUT 112, only an 8-bit code C0 can be obtained in a flat block, and two 4-bit codes C1 and C2 can be obtained in an edge block. Become. As described above, the color image data is obtained by compressing a 1-bit code indicating whether a predetermined pixel block of the color image data is a flat block or an edge block, and, for example, an 8-bit code and L * obtained by encoding the pixel block. The image data is compressed into the image data represented by the multi-bit code. The pixel configuration of the block and the method of selecting the chromaticity signal in this embodiment are completely arbitrary. Also, the input signal does not need to be R, G, B signals.
A signal obtained by A / D conversion of the SC YIQ composite signal may be used as it is, or I and Q may be handled as chromaticity signals. As described above, according to the present embodiment,
Efficient compression encoding can be performed with a fixed data length regardless of the flat part and edge part of the image.Especially in the flat part, the chromaticity signal can be encoded with a sufficient number of bits, so that a pseudo contour part does not occur. In addition, there is an effect that color smearing at the edge portion can be prevented. As described above, according to the present invention, the first encoding mode in which the chromaticity signal is divided into predetermined regions and each of them is encoded, and this predetermined region is further divided into a plurality of partial regions. When there is an encoding unit for selectively operating the second encoding mode for dividing and encoding each, the plurality of partial areas are
In order to perform efficient encoding considering the visual characteristics of humans, it is determined based on the content of the luminance signal whose changes are easily recognized by human eyes, not the chromaticity signal whose change is not easily recognized by human eyes. An optimal partial area can be set. Further, since the control is performed such that each code amount when each of the partial regions is coded in the second coding mode is smaller than the code amount when the predetermined region is coded in the first coding mode. In addition, efficient encoding that generates a code amount commensurate with the size of the encoding area can be performed. [0034]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of an image processing apparatus according to an embodiment. FIGS. 2A and 2B are diagrams for explaining the correlation between a luminance signal and a chromaticity signal at an edge portion. FIG. 2A shows the distribution of L *, and FIG. It is indicated by shading in pixel block units. FIG. 3 is a diagram illustrating an example of a bit configuration after encoding of an edge block and a flat block. FIG. 4 is a diagram illustrating an example of a data format conversion circuit. FIG. 5 is a diagram illustrating a distribution of a chromaticity signal of a reflection original. FIG. 6 is a diagram illustrating an example of area division. [Explanation of Code] 31 Code of flat block 32 Code of edge block 101 Color original 104 Color separation sensor 105 Sampling hold circuit (S / H) 106 A / D converter 107 Color separation circuit 108 Edge detection unit 109 Selectors 110, 111 Averaging circuit 112 LUT 115, 116 Latch circuit 117 Compression circuit 120 Area determination unit 121 An average calculation unit.

Claims (1)

(57) [Claims] And generating means for generating a luminance signal and chrominance signal from a color image signal, the chroma signal, a first encoding mode for each coded is divided into a predetermined area, the contents of said predetermined area said luminance signal wherein the second coding mode for further divided and each coded smaller plurality of partial areas a predetermined region, a coding means for operating selectively, the predetermined region by the first encoding mode determined based on And control means for controlling each code amount when each of the partial areas is coded in the second coding mode to be smaller than the code amount when coded. Color image processing apparatus. 2. The color encoding apparatus according to claim 1, wherein the encoding unit selectively operates the first encoding mode and the second encoding mode based on a chromaticity signal generated by the generation unit. Image processing device.