JPH01112867A - Encoding device for color picture information - Google Patents

Abstract

PURPOSE:To prevent the deterioration of a reproduced picture by encoding an original color picture as assuming a mean color to be the color information of the representative color of a block when the original color picture is a dot pictures. CONSTITUTION:A signal converter 1 separates a luminance information and a chrominance information from the color information of each picture element, and inputs them in a dot detector 2. A chrominance encoder 3 inputs the in- block picture information of L*, a*, b* and a dot detection signal, and encodes the chrominance information a*, b*. When the block intended to encode is a dot original, a color edge is not treated as the color edge even if it exists in the block. Therefore, the chrominance encoder 3 obtains the mean value of a* and b* in the block by performing sub-sampling on an a*, b* plane, and smooths it.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an encoding device for color image information, and particularly to an encoding device for encoding color image information of a halftone image.

[Prior Art] Conventionally, in an encoding device that encodes color image information in units of mxn blocks, extreme color differences exist within blocks (hereinafter, such color differences are referred to as color edges). In this case, in order to preserve this edge part, the inside of the block was encoded with color information of multiple colors. Since this color information is encoded by fixed-length encoding on a block-by-block basis, the color information of each of the plurality of colors is encoded with a smaller amount of information than the color information of the negative color.

On the other hand, a method is known in which color image information is separated into color information and brightness information and encoded separately. In this method, information about the structure of color information within a block generally depends on brightness information.

[Problems to be Solved by the Invention] When encoding a color halftone dot document using these encoding devices, a high frequency halftone dot structure within a block is changed to a low frequency structure. .

Furthermore, there are color edges within the block. In this case, in the case of high-frequency halftone dots, human vision does not perceive the halftone dot structure and perceives it as a mixture of multiple colors, but the encoded, decoded, and reproduced image is Since decoding is performed using multiple colors that have a structure and a small amount of information, the human eye perceives a roughly painted structure, giving a sense of roughness, especially in flat highlight areas.

[Means for Solving the Problems] The present invention was proposed in order to eliminate the drawbacks of the above-mentioned conventional example, and aims at high efficiency encoding and prevents deterioration of reproduced images of color halftone images. The purpose of this invention is to propose a possible encoding device for color image information.

The configuration of such an encoding device is such that the encoding device encodes color image information in units of blocks each consisting of a plurality of pixels, and includes a determining means for determining that the original color image is a halftone image, and a determining means for determining that the original color image is a halftone image, and a determining means for determining that the original color image is a halftone image. The block is characterized by comprising a calculating means for calculating an average color, and an encoding means for encoding the average color as color information of a representative color of a block when the determining means determines that the block is a halftone dot. shall be.

[Operation] According to this configuration, edges are easily picked up in the halftone image, but
When the original image is determined to be a halftone dot by the determining means, the pixel block has the average color of the entire block as the representative color of the block, so that no image quality deterioration occurs.

[Examples] Examples according to the present invention will be described in detail below with reference to the accompanying drawings.

<First Embodiment> FIG. 1 is a block diagram of an encoding device according to a first embodiment of the present invention. This encoding device inputs color image information block by block, performs encoding for each block, and outputs the encoded code. 2 is a signal converter for separating the information into brightness information (hereinafter referred to as brightness information) and color information (hereinafter referred to as chromaticity information), and 2 detects whether a block is a halftone image or not. This is a halftone dot detector. 3 is a chromaticity encoder that encodes the chromaticity information separated by the signal converter 1, and 4 is a brightness encoder that encodes the brightness information that is also separated. Reference numeral 5 denotes a combiner that combines the output encoded by the chromaticity encoder 3 and the output encoded by the brightness encoder 4 as one code and outputs the result.

The signal converter 1 receives the input color pixel block and separates the color information of each pixel into lightness information and chromaticity information. Generally, the input color image information is red (R).
It consists of three primary colors: , green (G), and blue CB).

As an information system separated into brightness information and chromaticity information, 19
76CIE L'a"b' space, 1976CIE L"
Y, I, Q for u”v” space or color TV signals
Space etc. are common. In this first embodiment, for convenience of explanation, the 1976 CIE a"b"b" space is used. In this space, L" represents lightness, a"b" represents chromaticity, R, G.

Conversion from B to L” a” b” is generally performed using RlG
, after converting from B to the XYZ color system, L″a′ b”
Convert to space. At this time, the conversion to the XYZ color system is -
This is done by the following matrix operation. Also Lm am
Conversion to b* space is expressed by the following equation that includes a local power term.

Y I/3 L"=116 (-) -18 Y.

Xo Y.

Yo Z.

... (1) (However, Xo, Yo, Zo are constants) In fact,
The signal converter 1 can be configured with a read-only memory (hereinafter abbreviated as ROM).

The image information of chromaticity and brightness thus converted is input to the halftone dot detector 2. Here, it is determined whether this pixel block is a halftone image. Various methods for detecting halftone dots have been reported. - For example, there is a method of determining and detecting statistics within a block. There is also a method of analyzing and detecting the frequency of a halftone dot, or a method of comprehensively determining a halftone dot using not only information from within the block but also information from adjacent blocks. Furthermore, the operator may operate a mode switch for setting the document to be halftone, and input from this switch may be used.

Information as to whether the image is a halftone dot image detected in this way is input to the chromaticity encoder 3 as a halftone detection signal.

The chromaticity encoder 3 inputs the intra-block image information of Lll, am, bg and the above-mentioned halftone dot detection signal, and
The chromaticity information a'', b' is encoded.

FIG. 2 is a block diagram of the chromaticity encoder 3. Signal converter 1
The chromaticity information of a' and b'' inputted from the terminals 12 and 13 is inputted to the separator 15. In this embodiment, when a color edge exists, for convenience of explanation, the chromaticity information of a' and b'' is inputted to the separator 15. It shall be.

The separator 15 separates the maximum values a', 1x, b'', II8 and the minimum value a''1 of a*, b* in the block, respectively.
lln+b"main. These maximum values and minimum values are input to the color edge determiner 16. First,
The color edge determiner 16 calculates the difference between the input maximum value and minimum value, that is, (a ``waax-a ''m1n) +
(1) Compare aaax-b ``win) and threshold values T, Tb, respectively, and if Ta < (a ``Pax-a ''+aln) or Tb <(b''waax-b''5ln) ,
It is determined that a color edge exists within the block. If not, it is determined that there is no color edge.

The relationship between the halftone original, color edges, and encoding is as follows. That is, when it is determined that the block to be encoded is a halftone original based on the halftone detection signal from the halftone dot detector 2 inputted from the terminal 11, even if a color edge exists in the block, This is not treated as a color edge. For this purpose, the color edge determination device 16 outputs a color edge determination signal to a selector 20, which will be described later. Therefore, this color edge determination signal becomes “1” only at halftone dots.
(Taku(a"s+ax-a"m1n)"Tbku(b
"+max-1)"+aln)).

The smoothing encoder 17 performs encoding of chromaticity information when there is no color two-way within the block. In this example,
For simplicity of explanation, the smoothing encoder 17 is
, b'' plane to obtain the average value x, T3 of am and bm within the block, thereby performing smoothing.

The edge encoder 18 uses the intra-block information input from the terminal 14 to first divide the pixels within the block into 2 pixels.
Separate into two groups. These groups are a group of pixels with an L'' value larger than the average value of Ll in the block, and a group of pixels with a value smaller than the average value.Then, the pixels in each of these groups are a
″.

The average value of the information of "b" on the a"b" plane is obtained by sampling, and the average value of these two sets (#1
10, S and #2M, T3) are the chromaticity information when a color edge exists in the block. These chromaticity information (#1
and #2, 5) are combined in the synthesizer 19 to have the same amount of information as the output of the smoothing encoder 1f.

Either the output of the smoothing encoder 17 or the output of the synthesizer 19 is selected by the selector 20 according to the color edge determination signal output from the color edge determination unit 16. That is, this selector 20 selects the output of the smoothing encoder 17 when the block is a halftone image or when there are no color edges in the block; If present, edge encoder 18
Select the output obtained from the synthesizer 19 from the terminal 2
Output from 2. Further, the presence or absence of a color edge is outputted from the terminal 21 as a color edge determination signal.

The bean 1j1L cell brightness information encoder 4 converts the brightness information into, for example, Mi
Block encoding proposed by tchell and Atsugi, vector quantization with less visual deterioration of structural information within a block, transform encoding, etc. are performed.

Tachizume 1 Brightness information L'' and chromaticity information a'' obtained in this way,
The encoded result of b', ie, [and (i, E) or (#1 and #2, 100, '5), and the color edge determination signal are combined by a combiner 5. The combined format will be as shown in Figure 3A if no color edges are present, and will have a code as shown in Figure 3B if color edges are present.

<Second example> FIG. 4 is a block diagram of a second embodiment according to the present invention.

In this second embodiment as well, for convenience of explanation, the L"a"b1 system is used as the image information information system, as in the first embodiment.

In FIG. 4, 31 is a signal converter for obtaining lightness information L'' and chromaticity information a'', b'' from the color image information in the block. This is a category classifier for classifying into categories whose rough structure is a set of similar items based on feature amounts.

Further, in this second embodiment, a method of using Hadamard transform, which is an orthogonal transform, and vector quantizing the result of the Hadamard transform is used as a method of encoding the lightness information L''.A Hadamard transformer 33 for that purpose is used. For example, input 4x4 block information, convert it to 1x1
6 matrix (or vector) X, and using the Hadamard transformation matrix H1ll, the transformation result Y
get.

Y=-HI 6X To Jl The Hadamard transformation matrix used here is as follows.

[Margins below] The brightness values thus subjected to Hadamard transformation are subjected to the above-mentioned vector quantization based on each category value obtained from the category classifier to obtain an encoding result.

5. The aLi north chromaticity information is input to the chromaticity information encoder 35, where it is encoded.

FIG. 5 is a block diagram of the chromaticity encoder 35. Terminal 42.
The a"b" information inputted from 43 is inputted to separator 46. These a", b" are the separator 1 of the first embodiment.
5, the maximum value of pixels in the block a ”@aK+
1) "l1laX and minimum value a" + ai□b "1n
and seek. These maximum values and minimum values are input to the color edge determiner 47. Similarly to the color edge determining device 16 of the first embodiment, this color edge determining device also calculates the difference between the maximum value and the minimum value, that is, (a″aall-a ”11111
) + (b ``maK-b ''1lln) and threshold T,
, Tb, respectively, and if Ta<(a"amx-a'5in) or Tb<(b", □-b"5in), it is determined that a color edge exists within the block. If not, it is determined that there is no color edge.

In the first embodiment, a block determined to be a halftone dot by the halftone detection signal is treated as having no color edge even if a color edge exists. If the general structure of this block is determined to be a category characterized by a halftone structure based on the category classification signal, even if there are color edges in the block,
This is not treated as a color edge.

The output of this color edge determiner 47 is input to an AND circuit 48. The other input of this AND circuit 48 comes from the vector quantizer 34 via a terminal 44. This is due to the following reasons.

That is, in the vector quantizer 34, brightness edges within a block may disappear. When this edge disappears, the color edge determination unit 47 determines that a color edge exists, and if the image is reproduced based on that determination,
It becomes impossible to paint separately using a plurality of colors, for example, two colors, using differences in brightness.

In such a case, deterioration in the image quality of the reproduced image will be smaller if the image is reproduced using the average color of these two colors than when the image is reproduced using either one of the two colors.

Therefore, the chromaticity encoder 35 receives the edge erasure signal from the vector quantizer 34.

That is, when it is determined that the brightness edge has disappeared due to vector quantization, the edge disappearance signal (
! "1") is input, and the AND circuit 48 calculates the AND with the chromaticity edge determination signal. Therefore, if there is no brightness edge, there is no chromaticity edge. That is, the output of the AND circuit 48 (color edge determination signal) becomes "1" when (edge exists) and (edge disappears).

Similar to the smoothing encoder 17 described above, the smoothing encoder 49 converts the average value of a* b* in the block into a*
Encoding is performed by subsampling on the b* plane.

In addition, the edge encoder 50 receives L8 input from the terminal 45.
Separate the pixels in the block into two groups using the intra-block information of encoded by subsampling on the synthesizer 5
1 combines these two codes. These smoothing,
The edge encoding 9 and the like are the same as in the first embodiment.

The encoded results of these two pieces of chromaticity information are combined into an AND circuit 4.
Switching is performed in the same way as the selector 20 described above based on the color edge determination signal obtained by 8.

In this way, the encoded result of the chromaticity information is output from the terminal 54, and the color edge determination signal is output from the terminal 53.

The synthesizer 36 combines the encoding results of the brightness information and chromaticity information obtained in this way, the color edge determination signal and the category classification signal, and when there is no color edge, the If a color edge exists in a code as shown in the figure, it is synthesized into a code as shown in FIG. 6B.

<Decoding> To decode the color information of each code in the first and second embodiments described above, first, refer to the color edge determination signal and determine whether the chromaticity information is one color or two colors.
Determine the color. If there is no color edge, the average color a”
If there is a 0-color edge to be reproduced by ``b'' and decoded L8, pixels with Ll larger than the average value are reproduced from #1 chromaticity information a''b'', and pixels smaller than the average value are reproduced from #2 A reproduced image is obtained by reproducing with a*bm of chromaticity information.

Modified example> In the first embodiment, as shown in FIG. 7, the halftone detection signal is input not only to the chromaticity information encoder 3 but also to the brightness information encoder 4 to smooth the brightness information. Of course, it is also possible to perform encoding after processing.

In the second embodiment, the color edge determination signal is not required in the category ``-'' indicating the halftone dot structure. Therefore, this is deleted and a larger amount of information is allocated to encoding brightness information, for example, as shown in FIG. This is because halftone dots have complex brightness information, so by allocating a large amount of information to the encoding of brightness information, deterioration in image quality during playback is prevented.

Further, the encoding method of the brightness information in the two embodiments described above is not particularly limited, and the encoding method of the chromaticity information is also not limited to this, although subsampling of the average value is used in the present embodiment. The brightness and chromaticity are not limited to these as long as they are similar.

[Effects of the Invention] As explained above, according to the color image information encoding device of the present invention, deterioration of the reproduced image is reduced by encoding the chromaticity of the halftone original using the average color within the block. In addition to eliminating the color edge determination signal within the fixed length code, there is an effect that highly efficient encoding can be performed.

[Brief explanation of the drawing]

Figure 1 is a block diagram of the first embodiment, Figure 2 is a block diagram of the chromaticity encoder of the first embodiment, Figures 3A and 3B are code diagrams of the first embodiment, and Figure 4. is a block diagram of the second embodiment, FIG. 5 is a configuration diagram of a chromaticity encoder in the second embodiment, FIGS. 6A and 6B are code configuration diagrams of the second embodiment, and FIG. FIG. 8 is a block diagram showing a modification of the first embodiment, and FIG. 8 is a code diagram of a modification of the second embodiment. In the figure, 1.31...signal converter, 2...halftone detector, 3.35...chromaticity encoder, 4...brightness encoder, 5.19, 35, 51 ．． ...Synthesizer, 15.46・
...Separator, 16.47...Color edge determiner, 17.
49... Smoothing encoder, 18.50... Edge encoder, 20.52... Selector, 32... Category classifier, 33... Hadamard transformer, 34... Vector Quantizer, 48...AND circuit. Patent applicant Canon Co., Ltd.

Claims (5)

[Claims] (1) In an encoding device that encodes color image information in units of blocks each consisting of a plurality of pixels, there is a determining means for determining that the original color image is a halftone image, and calculating the average color of pixels within the block. An image characterized by comprising: a calculation means; and an encoding means for encoding the block with the average color as color information of a representative color of the block when it is determined to be a halftone dot by the discrimination means. Information encoding device. (2) The encoding means includes: means for detecting a color edge within a block; means for, upon detecting a color edge, representing each group within the block divided by the edge with a plurality of colors; Claims characterized in that, even if the detecting means detects that a color edge exists, when the discriminating means detects a halftone dot, suppressing the means for representing with a plurality of colors. The image information encoding device according to item 1. (3) When the calculation means separates brightness information and color information from the color image information in the block and encodes the brightness information and the color information separately, if the original is a halftone printed matter, 2. The image information encoding apparatus according to claim 1, wherein only color information is encoded as information obtained by smoothing color information. (4) When separating brightness information and color information from the color image information in the block and encoding the brightness information and color information separately, if the document is a halftone print, the color information is 2. The image information encoding apparatus according to claim 1, wherein the brightness information is smoothed from the flat color information in the block and is encoded. (5) The image according to claim 1, characterized in that when the structural information in the block becomes completely flat as a result of encoding, the color information is encoded as the average color in the block. Information encoding device.