JP4116193B2 - Color image processing apparatus and color image processing method - Google Patents

Description

[0001]
The present invention provides a color image equipped with an image area separation device for reproducing a document with high image quality in a color image reproduction device such as a one-drum type color copying machine or color facsimile equipped with an image data compression function. The present invention relates to a processing apparatus and a color image processing method .
[0002]
[Prior art]
In a color copying machine, four-color data of Y (yellow), M (magenta), C (cyan), and K (black) is created from the read document data, and four-color planes are sequentially created and superimposed. Reproduces a full color image. At that time, generally, image area separation is performed, and moire removal, halftone dot removal, edge enhancement, and the like are performed by a method that is different for each region. In general, there are a halftone dot region, an edge region, a chromatic achromatic region, and the like as regions separated by such image region separation.
[0003]
However, in recent years, a read color image is once stored in a page memory for high-speed processing and editing processing. For example, in a one-drum type color copying machine, an image may be read once into a memory and sequentially read when a CMYK plate is produced.
[0004]
However, the amount of color image information is increasing as the resolution of the image is increased. If the read image signal is directly stored in the memory, the necessary memory becomes enormous. For example, when the resolution is 600 dpi, the gradation is 24 bits / pixel, and the paper size is A3, the capacity is 256 megabytes. Therefore, compression encoding is often performed to reduce memory costs.
[0005]
[Problems to be solved by the invention]
However, if image region separation is performed after compression encoding, the image is generally deteriorated by compression encoding. Therefore, there is a problem that the region is not correctly determined by conventional image region separation for uncompressed images. Such inconvenience is large in the separation of regions where local changes are severe, particularly the halftone dot separation.
[0006]
This is AC component information of image information in a general compression method (the AC component here refers not only to the AC component obtained when frequency conversion is performed, but also general structural information of an image. For example, when the map bit in GBTC (the coefficient for quantizing each pixel between the maximum value and the minimum value in the block) is reduced, the halftone dot region has many complex high-frequency components, and these complex This is due to the fact that reproducible alternating current components are not compatible with effective compression coding. In addition to the halftone dot region, the edge region and the like are regions that are difficult to separate by deleting the AC component.
[0007]
On the other hand, if image area separation is performed before compression coding, all the separation results must be stored in the memory, which increases the separation memory.
In Japanese Patent Laid-Open No. 5-145551, separation information obtained at the time of the first scan is stored in a memory in order to prevent image quality deterioration due to a difference in separation results between scans for a four-time scanner. A technique has been proposed in which the data is stored in the memory and used as a separation result in common for each scan.
[0008]
An object of the present invention is to prevent a change in separation result due to compression encoding, unlike the above-described conventional example that prevents variation in separation results between scans.
[0009]
[Means for Solving the Problems]
In order to achieve the above object, according to the first aspect of the present invention, an image input unit that reads an original in one scan and inputs image data, and compression-encodes the image data input by the image input unit. Compression means; storage means for storing image data compressed and encoded by the compression means; decompression means for decompressing image data stored in the storage means; and image data input by the image input means. First separation means for obtaining edge separation information and halftone separation information of the image data from the image data before being compressed and encoded by the compression means; and color separation from the image data decompressed by the decompression means second segmentation means for obtaining information, on the basis of the edge separation information, halftone separation information and said color separation information, and region determination means for performing area determination, Based on the determination result of the serial area determining means, a color image processing apparatus characterized by having an image processing unit that performs image processing on the decompressed image data by the decompression means.
[0011]
In order to achieve the above object, the invention described in claim 2 includes an image input step for obtaining image data by reading a document by one scan of the image input device, and image data obtained by the image input step. Obtained by the compression step for compression encoding, the storage step for storing the image data compressed and encoded by the compression step, the expansion step for expanding the image data stored in the storage step, and the image input step A first region separation step for obtaining edge separation information and halftone separation information of the image data from the image data before being compressed and encoded by the compression step; and the image decompressed by the decompression means a second segmentation step of obtaining color separation information from the data, on the basis of the edge separation information, halftone separation information and said color separation information, region-size performing area determination A step, based on the determination result of the region determination step, a color image processing method characterized by having an image processing step of performing image processing on the decompressed image data by the decompression process.
[0012]
In order to provide a low-cost and high-quality color image processing apparatus by reducing the amount of information using compression technology and preventing erroneous separation due to image area separation in an image after compression encoding, In the present invention, as described above, with regard to halftone dot separation and edge separation, in which erroneous separation due to compression coding is significant, the separation result is obtained from an image before compression coding, and this is stored in a memory, thereby degrading image quality. It is something to be suppressed.
[0013]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.
FIG. 1 is a block diagram of a color image processing apparatus showing an embodiment of the present invention.
This color image processing apparatus includes an image input device 101 such as a scanner, a compression / decompression circuit 102 for entropy-encoding data, a page memory 103 for holding compressed data, and a signal having a linear reflectance and converting it to a signal having a linear density. A log conversion circuit 104, a filter circuit 105 including a smoothing filter, a color correction circuit 106 for matching colors, and a UCR circuit 107 for subtracting K signals from each color material signal are provided.
[0014]
Similarly, a gradation processing circuit 108 for expressing halftones using a dither for a picture, an image area separation circuit 109 for separating an image into each area, an image output device 110 such as a printer, and a halftone dot from a signal before compression. Halftone separation unit 111 for separating only the region, halftone separation result memory 112 for storing the result of halftone separation, edge separation unit 113 for separating only the edge region from the signal before compression, edge separation The edge separation result memory 114 is stored for storing the results.
[0015]
Digital image data (8-bit signal) read from the image input device 101 is sent to the edge separation unit 113 and the halftone dot separation unit 111 and also sent to the compression / decompression circuit 102 as an RGB signal.
[0016]
The edge separation unit 113 is a circuit that detects an edge region from a document based on a signal immediately after being read. This specific configuration is described in, for example, a paper “Image Area Separation Method of Mixed Character / Picture (Half Dots and Photos)” (Electronic Information and Communication Society Transactions Vol. J75-DI1No. 1pp39-47 1992.1). Used method.
[0017]
This method will be described in detail. Edge separation is performed after edge enhancement is performed on input image data, and then ternarization is performed using threshold values TH1, TH2 (TH1> TH2) (if the target pixel is larger than TH1, white pixel, TH2 If it is smaller, it is a black pixel), and the continuity between the ternarized black pixel and the white pixel is detected by pattern matching, and one or more black continuous pixels and white continuous pixels exist in a 5 × 5 pixel block The pixel of interest is determined to be an edge region, and the other pixels are determined to be non-edge regions (processing in units of pixels).
[0018]
A halftone dot separation unit 111 is a circuit that detects a halftone dot region in a document based on a signal immediately after being read. This specific configuration is described in, for example, a paper “Image Area Separation Method of Mixed Character / Picture (Half Dots and Photos)” (Electronic Information and Communication Society Transactions Vol. J75-DI1No. 1pp39-47 1992.1). Used method.
[0019]
This method will be described in detail. Focusing on the fact that the density change of the halftone dot area is significantly different from that of the character area, peak pixel detection, halftone dot area detection, halftone dot area correction is performed, and the halftone dot area is separated. Is. For example, in a 3 × 3 pixel block, the peak pixel is detected by a pair of pixels in which the density level L of the central pixel is higher or lower than that of all the surrounding pixels and is located on a diagonal line across L and the central pixel. When the density levels a and b of the pixels are | 2 × L−a−b |> TH3 (threshold), the center pixel is set as the peak pixel.
[0020]
For example, in the detection of a halftone dot region, if there are two or more blocks including peak pixels in four blocks each having 4 × 4 pixels as a unit, the target block is determined as a halftone dot candidate region, and other than that, a non-halftone dot is detected. It is determined as a candidate area. After judging the halftone / non-halftone dot candidate area, in the nine blocks centered on the target block, if more than 4 blocks are halftone dot candidate areas, the target block is set as the halftone dot area, otherwise the target block is selected. A non-halftone area is set (block unit processing).
[0021]
The compression / decompression circuit 102 first converts the image signal to reduce entropy and then compresses it. As a compression method, for example, fixed length compression is performed here. Specifically, the conversion unit localizes the statistical power of the image by, for example, color conversion and subband conversion, and reduces entropy. Here, YIQ conversion is used as an example of color conversion.
[0022]
Y = 0.3R + 0.59G + 0.11B
I = 0.74 (R−Y) −0.27 (B−Y)
Q = 0.48 (R−Y) +0.41 (B−Y)
[0023]
As an example of the subband transform, Haar wavelet transform is used as shown in FIG. This is a conversion method for decomposing pixel values a (upper left), b (upper right), c (lower left), and d (lower right) of a 2 × 2 pixel block into four coefficients LL, HL, LH, and HH. . Here, the characteristics of each coefficient will be outlined. The LL coefficient represents a DC component (average of four pixels) in the block, and the HL, LH, and HH coefficients represent an AC component. Further, among the AC components, the HL coefficient is information representing a horizontal edge, the LH coefficient is representing a vertical edge, and the HH coefficient is information representing an oblique edge.
[0024]
The compression of the coefficient after conversion is performed as follows, for example.
First, the amount of information is reduced by sub-sampling the I and Q coefficients. At this time, it is effective to perform adaptive sub-sampling such as performing different sub-sampling in the character part and the picture part in consideration of visual image quality. For the Y coefficient, the Haar wavelet transform described above is performed, and the obtained coefficient LL (8 bits), HL (9 bits), LH (9 bits), and HH (10 bits) are quantized to reduce the amount of information. Reduce.
[0025]
At this time, it is effective to perform adaptive quantization such as performing different quantization in the character part and the picture part in consideration of visual image quality, and vector quantization is used for the quantization of AC components. It is valid. Finally, the quantized value is embedded in one code (for example, 12 bits) and fixed-length compressed. In this case, when adaptive quantization or adaptive subsampling is performed, it is necessary to embed region information in the code.
[0026]
The compressed code is held in the page memory 103, and when the second version is created, the compressed code read from the page memory 103 is decoded in the reverse procedure of the compression, and decompressed image data R ', G' , B ′.
[0027]
The input data is sent to the compression / decompression circuit 102, and is also sent to the Log conversion circuit 104 at the same time when the K version is created. On the other hand, when the C, M, and Y versions are created, the data after compression / expansion is sent to the log conversion circuit 104. The log conversion circuit 104 performs table conversion to convert the signal characteristics from the reflectance space to the density space. Here, the output signal represents the ink amount.
[0028]
The filter circuit 105 is composed of various filters, and performs the most appropriate filter processing according to the area that the image area separation circuit 109 described later determines that the image signal belongs to. For example, smoothing filter processing is performed on the pattern portion, and edge enhancement filter processing is performed on the character portion.
[0029]
The color correction circuit 106 is a circuit for removing the turbidity component of the color separation filter in the scanner or the turbidity component of the ink, and generally includes a masking method, a tetrahedral interpolation method, a triangular prism interpolation method, and the like.
[0030]
Here, the masking method is used, and the conversion formula is shown below.
C = k11 * c + k12 * m + k13 * y + k14
M = k21 * c + k22 * m + k23 * y + k24
Y = k31 * c + k32 * m + k33 * y + k34
Here, the constants k11 to k34 are determined by experiments.
[0031]
The UCR circuit 107 switches processing as follows based on the area determination information from the image area separation circuit 109. That is, in the black character region, K is set to “through” and C = M = Y = 0. In the color character area, K, C, M, and Y are set to “Through”.
K ′ = 0.6 × K
C ′ = C−K ′
M ′ = M−K ′
Y '= Y-K'
And
[0032]
The gradation processing circuit 108 prepares two types of dither tables having different sizes. When the signal from the image area separation circuit 109 is a color character or a black character, the dither table is processed with a small size dither table (for example, 1 × 1). In other cases, processing is performed with a large dither table (for example, 2 × 2), and the processed data is sent to the image output device 110 and output.
[0033]
Next, the image area separation circuit 109 will be described. In the present embodiment, the image area separation circuit 109 includes a color separation circuit 121 and a determination circuit 122 as shown in FIG. The color separation circuit 121 includes, for example, the following two steps. That is, in the first step, max (cm, my, yc) of the target pixel is obtained, and when this value is larger than TH4 (threshold), the target pixel is set as a chromatic pixel. In the second step, the chromatic pixels are counted in the target block (4 × 4 pixels), and when the coefficient value is larger than a predetermined threshold TH5, the target block is set as a color block (block unit processing).
[0034]
The determination circuit 122 receives the result obtained from the color separation circuit 121, the halftone separation result stored in the halftone separation result memory 112, and the edge separation result (all on and off) stored in the edge separation result 114. The following determination is made. That is, a black character area signal is generated by edge separation (on), halftone separation (off), and color separation (off), and a color character area is generated by edge separation (on), halftone separation (off), and color separation (on). A signal is generated, and all others generate a picture area signal. These region signals are sent to the filter circuit 105, the UCR circuit 107, and the gradation processing circuit 108, and optimum processing is performed for each region as described above.
[0035]
In the above configuration, the present invention will be described below.
As described above, the compression coding method in the present embodiment is a method of performing sub-sampling for discarding all the high frequencies HL, LH, and HH for the I and Q signals, and vector quantizing the high frequency for the Y signal.
[0036]
By the way, when image information is reduced by compression coding, information that is most hardly noticeable in deterioration caused by the reduction is information on an AC component, particularly, an oblique edge. For example, if the direct current component is reduced, a very large deterioration such as a change in the color of the image occurs, and sometimes the original image cannot be recognized. Therefore, it is only possible to obtain a blurred image with a drop in the image quality. Further, even when the vertical edge and the horizontal edge are reduced among AC components, the edge portion of the character is chipped or blurred, and the deterioration is noticeable, while the deterioration of the oblique edge is difficult to see.
[0037]
Under such circumstances, when compressing an image, it is a common practice to reduce the amount of information by further reducing AC components, particularly HH coefficients representing oblique edges. In the case of a color image, the luminance information is more important than the color difference information. Therefore, the color difference information is greatly reduced compared to the luminance information. The compression method in the present embodiment is also in accordance with such a general method.
[0038]
Under such circumstances, generally, the diagonal edge of the color difference and the diagonal edge of the luminance (or the diagonal edge of each color when compression encoding without color conversion is used) are deteriorated by the compression encoding. In other words, the edge in the oblique direction of the image is lost due to the compression encoding.
[0039]
On the other hand, image area separation of halftone dots is generally difficult when a slanted edge is sagged. For example, the halftone dot separation method described above is “in the 3 × 3 pixel block, the density level L of the central pixel is higher or lower than that of all the surrounding pixels, and exists on a diagonal line with L and the central pixel in between. When the determination condition is that the density levels a and b of the two pixels are | 2 × L−a−b |> TH3 (threshold) for all four pairs, it is clearly | 2 × L−a−b | does not easily exceed TH3.
[0040]
As described above, since the halftone dot region is a region having a complex edge structure including many oblique edges, it is not possible to perform high-precision image region separation after performing efficient compression coding. very difficult.
[0041]
Therefore, as in the present invention, if the halftone dot region that is very difficult to separate after compression encoding is separated in advance before compression, and the result is stored in the separation memory 112, the above difficulty is solved, It is possible to perform image area separation with high accuracy while performing efficient compression.
[0042]
For the same reason, generally, the vertical edge and horizontal edge of color difference / luminance (or the vertical edge and horizontal edge of each color when using compression encoding without color conversion) are also deteriorated by compression encoding. Deterioration of vertical and horizontal edges is conspicuous in terms of image quality, so more information is generally assigned than diagonal edges, so it is not as distorted as diagonal edges. You will be drowned.
[0043]
On the other hand, the edge separation becomes difficult when the edge falls. For example, the above-described edge separation method is “after performing edge emphasis on input image data, it is ternarized using threshold values TH1, TH2 (TH1> TH2) (if the target pixel is larger than TH1, white pixel, TH2 If it is smaller, it is a black pixel), and the continuity between the ternarized black pixel and the white pixel is detected by pattern matching, and one or more black continuous pixels and white continuous pixels exist in a 5 × 5 pixel block This is done by determining the pixel of interest as an edge region and determining the other as a non-edge region. ”However, if an edge drift occurs, the result of ternarization changes, and it is determined as a black pixel or a white pixel. There will be fewer pixels.
[0044]
In this way, considering the situation on the compression side of actively reducing the AC component during compression encoding, it is possible to perform image region separation with high accuracy after performing efficient compression encoding. difficult.
[0045]
Therefore, if the edge region that is difficult to be separated after compression coding is separated in advance before compression as in the present invention and the result is stored in the separation memory 114, the above-mentioned difficulty is eliminated and efficient. It is possible to perform image area separation with high accuracy while performing compression.
[0046]
According to the present invention , the page memory can be reduced by using compression encoding, and the memory cost can be kept low. Further, image area separation is performed before compression coding, and the result is stored in a memory, so that deterioration in image quality due to erroneous separation due to compression coding can be suppressed. Furthermore, by performing only image area separation, which is difficult to separate with image data after compression coding, before compression coding, there is an advantage that it is not necessary to use a separation memory more than necessary.
[0047]
Further, according to the present invention, by performing halftone dot separation, which is difficult to separate with high accuracy in image data after compression coding, before compression coding, effective compression coding can be performed without worrying about erroneous separation. Therefore, the memory cost can be reduced and the image area separation can be performed with high accuracy.
[0048]
Further, according to the present invention, by performing edge separation, which is difficult to separate with high accuracy in image data after compression coding, before compression coding, effective compression coding can be performed without worrying about erroneous separation, The memory cost can be reduced and the image area separation can be performed with high accuracy.
[Brief description of the drawings]
FIG. 1 is a block diagram of a color image processing apparatus showing an embodiment of the present invention.
FIG. 2 is a diagram illustrating a Haar wavelet transform method.
FIG. 3 is a block diagram of an image area separation circuit.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 101 Image input device 102 Compression / decompression circuit 103 Page memory 104 Log conversion circuit 105 Filter circuit 106 Color correction circuit 107 UCR circuit 108 Gradation processing circuit 109 Image area separation circuit 110 Image output device 111 Halftone separation unit 112 Halftone separation result memory 113 Edge separation unit 114 Edge separation result memory

Claims (2)

Image input means for reading a document in one scan and inputting image data;
Compression means for compressing and encoding image data input by the image input means;
Storage means for storing image data compression-encoded by the compression means;
Decompression means for decompressing the image data stored in the storage means;
First region separation means for obtaining edge separation information and halftone separation information of image data from image data input by the image input means and before being compressed and encoded by the compression means;
Second region separation means for obtaining color separation information from the image data decompressed by the decompression means;
Area determination means for performing area determination based on the edge separation information, halftone dot separation information and the color separation information ;
Image processing means for performing image processing on the image data expanded by the expansion means based on the determination result of the area determination means;
A color image processing apparatus comprising: An image input step of obtaining image data by reading a document in one scan of the image input device;
A compression step of compressing and encoding the image data obtained by the image input step;
A storage step of storing the image data compressed and encoded by the compression step;
A decompression step of decompressing the image data stored in the storage step;
A first region separation step of obtaining edge separation information and halftone separation information of the image data from the image data obtained by the image input step before being compressed and encoded by the compression step;
A second region separation step of obtaining color separation information from the image data decompressed by the decompression means;
A region determination step for performing region determination based on the edge separation information, halftone dot separation information and the color separation information ;
An image processing step for performing image processing on the image data expanded by the expansion step based on the determination result of the region determination step;
A color image processing method comprising:

Images