JPH08336051A - Image processor - Google Patents

Abstract

PURPOSE: To perform more efficient compression and picture processing of picture data by discriminating the attributes of the belonging pictures of a block based on the value of a gradation width index and the average value information of brightness components and chromaticity components. CONSTITUTION: In an encoding/decoding processing part 604, after executing the encoding processing of a GBTC system, the attributes of the image are discriminated by respective block units and encoded data (the gradation width index, the average value information and code data) are compressed again based on a discriminated result. Compressed data obtained by compression for the second time are stored in a compressed image memory 610 and a flag for indicating the images is stored on a hard disk 614. In the encoding/decoding processing part 604, based on the value of the flag for indicating the attributes of the images, the re-compressed data are expanded to the encoded data. Then, the encoded data are decoded corresponding to the GBTC system and the lightness component L2* and the chromaticity components a2* and b2* of the image data are outputted.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to GBTC (Generalize
The present invention relates to an image processing device that compresses and encodes image information using a d Block Truncation Coding) method.

[0002]

2. Description of the Related Art In recent years, a GBTC method has been proposed as a method for compressing and expanding image data of a document. GBTC
In the method, image data of an original is extracted for each block of a predetermined pixel matrix, and for each block, an average value Q1 of data equal to or less than a parameter P1 defined by the data in the block and a parameter P2 (where P1 <P2 are satisfied). Based on the average value information LA obtained by dividing the sum of the average values Q4 of the data having the above values into two equal parts and the gradation width index LD which is the difference between the average values Q4 and Q1. The data of each pixel in the block is compression-encoded into code data φij obtained by quantizing the data to a gradation level smaller than the data within the gradation distribution range in the block. FIG. 1 is a diagram for explaining the flow of a general GBTC encoding process. In the GBTC method, as shown in (a), image data of a document image is extracted in units of 4 × 4 pixel blocks. The image data in the extracted 4 × 4 pixel block is subjected to GBTC encoding processing. By this GBTC method encoding process, 1 byte (= 8 bits) of data for each pixel × 16 pixels of image data (16 bytes, that is, 128 bits) is converted into a 1-byte layer as shown in (b). The adjustment index LD, the average value information LA of 1 byte, the 2-bit code data assigned by classifying the data of each pixel into four stages, and a total of 16 pixels totaling 6
It is encoded into byte (= 48 bits) data. As a result, the data amount of the image is compressed to 3/8.
(C) is a figure showing that the data amount of the encoded image data corresponds to 6 pixels of the image data before encoding. Decoding of the encoded data is executed by calculating 1-byte image data corresponding to each 2-bit encoded data based on the gradation width index LD and the average value information LA.

Pixels Xij in a 4 × 4 pixel block
(However, i and j are values of 1, 2, 3, and 4, respectively.) The image data is 256 gradation data of four kinds of values by the encoding process and the decoding process of the GBTC method. Is replaced by Here, the decrypted data is
When compared with the data of the original image, it clearly contains an error. However, the error is a level that is not noticeable due to human visual characteristics, and almost no deterioration in image quality is recognized in a natural image. In the GBTC method, the parameter Q is calculated from the gradation range index LD and the average value information LA included in the encoded data.
1 and Q4 can be determined. That is, the parameter P
A character image composed of a black portion of 1 or less and a white portion of parameter P2 or more can be reproduced from encoded data. In the JPEG method in which Huffman coding is performed on the data obtained by DCT converting the image data, the compression rate of the data changes depending on the type of the original. That is, it is possible to achieve higher data compression for a certain original document than the GBTC method, but it is almost impossible for another original document to be compressed. Therefore, it is difficult to set the capacity of the memory provided in the image processing apparatus. However, in the GBTC method, data can be compressed at a constant compression rate. Therefore, there is an advantage that the capacity of the memory included in the image processing apparatus can be easily set.

In the image processing apparatus for executing the JPEG encoding process, the image data of the original is DCT-converted in units of blocks each having a predetermined pixel matrix. D
The CT-converted data has the average value information LA and the gradation width index LD obtained by the GBTC encoding process, and the average value and the gradation width of the data of each pixel forming the block like the code data φij. The data is not reflected. Therefore, it is very difficult to modify the density and color balance of an image reproduced by processing the DCT-converted data. Therefore, determine the type of original,
In order to edit / process the image reproduced on the sheet based on this determination result, it is necessary to perform these processes before the encoding of the image data or after the decoding.

[0005]

DISCLOSURE OF THE INVENTION As described above,
The GBTC image data encoding always shows a constant data compression rate (3/8). However, this compression rate is
When compared with the encoding of image data of the PEG method, the value is not so high. Further, in the image processing apparatus which executes the JPEG encoding process, the determination of the type of the original document is executed before the image data is encoded or after the image data is decoded. Further, AE processing for optimizing the density distribution of an image and image editing / processing such as color conversion and frame editing have also been executed before encoding image data or after decoding. In this case, a hardware circuit that performs high-speed calculation to determine the image attribute and a large-capacity memory that stores all image data are required, resulting in high cost.

An object of the present invention is to provide an image processing apparatus which executes more efficient image data compression and image processing.

[0007]

In a first image processing apparatus of the present invention, a data conversion processing unit for converting RGB image data of a document into lightness component data and chromaticity component data, and a lightness component and chromaticity. The component data is divided into blocks each consisting of a predetermined pixel matrix, and for each block,
The average value Q1 and the parameter P2 of the data of the value less than the parameter P1 determined by the data in the block (however, P
The average value Q of data having a value of 1 or more (having a relationship of <P2)
Based on the average value information obtained by dividing the sum of 4 into two, and the gradation width index which is the difference between the average value Q4 and the average value Q1, the data of each pixel in the block is converted into the floor of the block. An encoding processing unit that encodes encoded data obtained by quantizing at a gradation level smaller than the above-mentioned data within the range of the tone distribution, and an average of the lightness component and the chromaticity component with respect to the encoded block. An attribute discrimination processing unit that discriminates the attribute of the image to which the block belongs based on the value information, the gradation width index, and the value of the code data, and the average value information, the gradation width index, the code data, and the attribute judgment for each block. A storage unit that stores the result and a decoding processing unit that decodes the code data in block units based on the average value information stored in the storage unit and the gradation range index are provided.

Further, in the second image processing apparatus of the present invention,
A data conversion processing unit for converting RGB image data of an original into lightness component data and chromaticity component data, and the lightness component and chromaticity component data are divided into blocks each having a predetermined pixel matrix, and each block is divided into blocks. In addition, the average value Q1 and the parameter P2 (where P1 <P2
Based on the average value information obtained by dividing the sum of the average values Q4 of the data having the above values into two equal parts and the gradation width index which is the difference between the average values Q4 and Q1. An encoding processing unit that encodes the data of each pixel in the block into encoded data obtained by quantizing the data at a gradation level smaller than the data within the range of the gradation distribution in the block; With respect to the applied block, an attribute discrimination processing unit that discriminates the image attribute to which the block belongs based on the average value information of the lightness component and the chromaticity component, the gradation width index, and the value of the code data, and the lightness component and the chromaticity component. Average value information, gradation range exponent, a compression processing unit that performs a compression process that reduces data that can always be identified based on the attribute of the determined image among the code data, the data after the compression process, Attribute determination A storage unit that stores the results in association with each other, a compressed data decompression that reads the compressed data and the attribute determination result stored in the storage unit, and restores the data reduced in the compression process based on the attribute determination result. A processing unit, and a decoding processing unit that decodes coded data in blocks based on the average value information of the lightness component and chromaticity component expanded by the data expansion processing unit and the gradation range index.

Further, in the first and second image processing apparatuses, the attribute discrimination processing section may determine that the gradation width indices of the lightness component and the chromaticity component of the block to be discriminated are all below a predetermined value. When the block process determines that the block belongs to a solid image, and if even one value of the tone width index of the lightness component and the chromaticity component has a value that is equal to or more than the predetermined value, the block image has two or more floors. If the determination process of determining that the block belongs to a toned image and the average value information of the chromaticity components of the block to be determined are both equal to or less than a predetermined value different from the above, the block is determined to belong to a monochrome image. If the value of the chromaticity component average value information is equal to or more than a predetermined value different from the above, the block is determined to belong to a color image, and the brightness of the block to be determined is determined. Assigned to ingredients If the value of the encoded data is polarized, it is determined that the block belongs to a binary image. If the value is not polarized, the block is determined to belong to a multi-valued image. It is desirable to execute at least one discriminating process. In addition, before decoding the coded data in block units in the encoding processing unit, according to the attribute of the block, a histogram of brightness components of all blocks of the attribute, or a histogram of average value information of the brightness component and the chromaticity component is displayed. It is desirable to include average value information conversion processing means for converting the value of the average value information of each block into an appropriate value based on the obtained histogram. Further, from the storage unit, the average value information of the lightness component and the chromaticity component, the gradation range index, the code data and the attribute determination result are read in block units, and based on the read attribute determination results, the average value information of the corresponding component, It is desirable to include an edit / modification processing unit that changes the gradation width index and the value of the code data to predetermined values.

[0010]

In the first image processing apparatus, the attribute discrimination processing section is
The attribute of the image to which the block belongs is determined based on the average value information of the lightness component and the chromaticity component and the value of the gradation width index. Thereby, the attribute of an image can be determined without using the image data before the encoding process and after the decoding.
Further, since the attribute is determined for each block, it is possible to obtain more detailed information.

Further, in the second image processing apparatus, the compression processing unit selects from among the data coded by the coding processing unit.
The average value information of the lightness component or chromaticity component, the gradation width index, and the value of the code data that can be always specified by the attribute of the image of the block are reduced. The storage unit stores the data after the compression processing and the attribute determination result. This allows
The compression rate of image data can be significantly improved.

In the first and second image processing apparatuses, the attribute discrimination processing section may be arranged such that when the gradation width indices of the lightness component and chromaticity component of the block to be discriminated are all below a predetermined value, When the block processing determines that the block belongs to a solid image, and when even one value of the tone width index of the lightness component and the chromaticity component has a value equal to or more than the predetermined value, the block image has two or more floors. If the determination process of determining that the block belongs to a toned image and the average value information of the chromaticity components of the block to be determined are both equal to or less than a predetermined value different from the above, it is determined that the block belongs to a monochrome image. If the value of the chromaticity component average value information is equal to or greater than a predetermined value other than the above, the block is determined to belong to a color image, and the brightness of the block to be determined is determined. Assigned to ingredients If the value of the encoded data is polarized, it is determined that the block belongs to a binary image, and if the value is not polarized, the block is determined to belong to a multivalued image. If at least one discrimination process is executed, the attribute of the image to which the block belongs can be discriminated more effectively. That is, the block image is a monochrome solid image, a color solid image, a monochrome binary image, a monochrome multi-valued image, or a full-color image, depending on the combination of the above-described determination processes. Can be determined. Further, in the encoding processing unit, before decoding the code data in block units, according to the attribute of the block, the brightness component of all blocks of the attribute, or the average value information of the brightness component and the chromaticity component A histogram is obtained, and the average value information conversion processing means for converting the value of the average value information of each block into an appropriate value is provided based on the obtained minimum value, maximum value and average value of the histogram, or from the storage unit, in block units. Read the average value information of the lightness component and chromaticity component, the gradation width index, the code data and the attribute discrimination result, and based on the read attribute discrimination result, the average value information of the corresponding component, the gradation width index and the code data value. By providing an edit / modification processing unit for changing the value to a predetermined value, AE processing and color conversion for optimizing the density distribution of the original without processing the data before encoding or after decoding It is possible to perform the editing / processing such as frame editing, it is possible to realize the phenomena of the storage unit and the processing time required for the process.

[0013]

EXAMPLE A digital color copying machine of this example is a GBT.
After the encoding method of the C method is executed, the attribute of the image is discriminated in each processing block unit, and the encoded data is further compressed based on the discrimination result. What are the attributes of the image above?
It refers to five types of black and white solid images, color solid images, black and white binary images, black and white multivalued images, and full color images. Also, GBTC
The tone width index LD and the average value information LA included in the data encoded by the method are used to execute an AE process for optimizing the density distribution of an image, and an image editing / processing process such as color conversion or frame editing. Hereinafter, regarding the image processing apparatus of the present embodiment,
The following description will be made with reference to the accompanying drawings. (1) Encoding of image data by GBTC method (2) Configuration of digital color copying machine (2-1) Configuration of digital color copying machine (2-2) Operation panel (3) Description of image processing (3-1) Main routine (3-2) Key input processing (3-3) Recompression processing (3-3-1) Attribute discrimination processing <3-3-1-1> Solid image discrimination processing <3-3-1-2> Color / monochrome discrimination processing <3-3-1-3> Binary / multi-value discrimination processing (3-3-2) Feature amount extraction processing <3-3-2-1> Monochrome solid image feature amount extraction processing <3 -3-2-2> Color solid image feature amount extraction process <3-3-2-3> Monochrome binary image feature amount extraction process <3-3-2-4> Monochrome multi-valued image feature amount extraction process <3-3-2-4>-3-2-5> Full color image feature extraction processing (3-3-3) Recompression processing <3-3-3-1> Monochrome solid image recompression processing <3-3-3-2> Color solid image Recompression process <3-3-3-3> Monochrome binary image recompression process <3-3-3-4> Monochrome multivalue image recompression process (3-4) Decompression process from recompression ( 3-4-1) Decompression process from recompression <3-4-1-1> Decompression process from monochrome solid image recompression <3-4-1-2> Decompression process from color solid image recompression <3 -4-1-3> Decompression processing from monochrome binary image recompression <3-4-1-4> Decompression processing from monochrome multivalued image recompression <3-4-1-5> From full color image recompression Expansion processing (3-4-2) AE processing <3-4-2-1> Monochrome multi-valued image AE processing <3-4-2-2> Full color image AE processing (3-4-3) Image editing / Processing <3-4-3-1> Black and white solid image editing / processing <3-4-3-2> Color solid image editing / processing <3-4-3-3> Monochrome binary image editing / processing Processing <3-4-3-4> Black-and-white multi-valued image editing / processing <3-4-3-5> Full-color image editing / processing

(1) Encoding of image data by GBTC system In the GBTC system, image data of an original is extracted for each block of a predetermined pixel matrix, and for each block, a parameter P1 or less defined by the data in the block is used. Average value information LA obtained by halving the sum of the average value Q1 of the data and the average value Q4 of the data having a value equal to or greater than the parameter P2 (where P1 <P2 is satisfied) and the average value Q4.
And the gradation range index LD which is the difference between the average value Q1 and
The data of each pixel in the block is compression-encoded into code data obtained by quantizing the data to a gradation level smaller than the data within the range of the gradation distribution in the block.
FIG. 1 shows the G executed by the digital color copying machine of this embodiment.
It is a figure which shows the flow of the encoding process of a BTC system. GBT
In the C method, as shown in FIG. 1A, image data of a document image is extracted in units of 4 × 4 pixel blocks. The image data in the extracted 4 × 4 pixel block is encoded by the method described below with reference to FIG.
Byte (= 8 bits) data × 16 pixel image data (16 bytes, that is, 128 bits) is converted into a 1-byte gradation range index LD and a 1-byte average value as shown in FIG. 1B. The information LA and the data of each pixel are coded into data of 2 bytes code data × 16 pixels, which is assigned by classifying into 4 stages, for a total of 6 bytes (= 48 bits).
As a result, the data amount is compressed to 3/8. Figure 1 (c)
FIG. 4 is a diagram illustrating that the data amount of encoded image data corresponds to 6 pixels of image data before encoding. Decoding of the encoded data is executed by calculating 1-byte image data corresponding to each 2-bit encoded data based on the gradation width index LD and the average value information LA.
It should be noted that in the present embodiment, the image data of the document is 4 × 4.
Extraction is performed in pixel block units, but is not limited to this and 3
It may be extracted in units of × 3 pixel blocks or 6 × 6 pixel blocks. In addition, 2 of each pixel in the block
The 56-gradation data is not limited to being encoded into 4-gradation code data, and may be encoded into 2-gradation or 8-gradation code data. As will be described in order, in the image processing apparatus of the present invention, the image attribute determination is performed for each block using the average value information LA and the gradation width index LD obtained from the parameters P1 and P2 defined by the data in the block. It is also characterized in that various image processing is executed based on the discrimination result.

FIGS. 2A and 2B are diagrams showing GBTC system encoding processing and decoding processing. FIG. 2A shows the relationship among the maximum value Lmax and the minimum value Lmin, the parameters P1 and P2, and the gradation width index LD.
From the image data extracted in 4 × 4 pixel block units, a predetermined feature amount required for encoding is obtained. The feature amount is obtained by the following calculation. First, the maximum value Lmax and the minimum value L of each 8-bit image data in the 4 × 4 pixel block
Detect min. Next, a parameter P1 obtained by adding 1/4 of the difference between the maximum value Lmax and the minimum value Lmin to the value of the minimum value Lmin, and 3/4 of the above difference to the value of the minimum value Lmin.
The parameter P2 is calculated by adding The parameters P1 and P2 are calculated by the following equations 1 and 2.

## EQU1 ## P1 = (Lmax + 3Lmin) / 4

## EQU2 ## P2 = (3Lmax + Lmin) / 4 Next, of the image data of each pixel, the average value Q1 of the image data of the pixels having the parameter P1 or less is obtained. Further, of the image data of each pixel, the average value Q4 of the image data of pixels having the parameter P2 or more is obtained. Obtained average values Q1 and Q4
Based on, average value information LA = (Q1 + Q4) / 2,
The gradation width index LD = Q4-Q1 is calculated. Next, "Number 3"
And the calculation of “Equation 4” are performed to determine the reference values L1 and L2.

## EQU3 ## L1 = LA-LD / 4

[Equation 4] L2 = LA + LD / 4 The reference values L1 and L2 are, together with the average value information LA,
It is used when encoding 1 byte (8 bits) of each pixel, that is, 256-gradation image data into 2-bit, that is, 4-gradation code data. Next, code allocation is performed. 2B, in the 4 × 4 pixel block, the i-th column (where i = 1, 2, 3, 4; the same applies hereinafter) and the j-th row (where j = 1). , 2, 3, 4, which is the same below), the value of the code data φij assigned according to the data value of the pixel Xij. More specifically, the pixel Xij
The 2-bit code data φij having the value shown in the following “Table 1” is assigned according to the value of

[Table 1] The data encoded by the GBTC method is composed of code data (16 × 2 bits) for 16 pixels, a gradation width index LD and average value information LA of 1 byte (8 bits) each.

As shown in FIG. 2C, when the encoded data is decoded, the gradation width index LD is set.
And average value information LA are used. That is, according to the value of the code data φij assigned to the pixel Xij on the j-th row of the i-th column, the data of Xij is set to the value of 2 shown in the following “Table 2”.
Replace with 56 gradation data.

[Table 2]

That is, in the GBTC system, the decoded 4
Pixels Xij in the × 4 pixel block (where i and j
Are values of 1, 2, 3, and 4, respectively. The image data of) is replaced with 256 gradation data of four kinds of values. Therefore, the decoded data obviously contains an error when compared with the data of the original image. However, the error is a level that is inconspicuous due to human visual characteristics,
Almost no deterioration of image quality is observed in natural images. GBTC
In the method, the parameters Q1 and Q4 are completely restored from the gradation width index LD and the average value LA included in the encoded data. Therefore, in the character image, if the black portion is the parameter P1 or less and the white portion is the parameter P2 or more, the character image can be completely restored. In the JPEG method in which encoding is performed using DCT conversion, the data compression rate changes depending on the type of document. That is, higher data compression than the GBTC method with less deterioration is realized for natural images, but it may be almost impossible to compress binary images, CG images, and character images. Therefore, it is difficult to set the capacity of the memory provided in the image processing apparatus. In the GBTC method, data can be compressed at a constant compression rate. Therefore, there is an advantage that the capacity of the memory included in the image processing apparatus can be easily set.

(2) Configuration of Digital Color Copying Machine (2-1) Configuration of Digital Color Copying Machine FIG. 3 is a sectional view of the digital color copying machine of this embodiment. The digital full-color copying machine is roughly divided into an image reading unit 100 that reads RGB image data of a document and a copying unit 200. In the image reading unit 100, a document placed on the platen glass 107 is exposed by the exposure lamp 101.
Is illuminated by. The reflected light of the original is reflected by the three mirrors 10.
It is guided to the lens 104 by 3a, 103b and 103c, and an image is formed by the CCD sensor 105. Exposure lamp 101
The mirror 103a is moved by the scanner motor 102 in the arrow direction (sub-scanning direction) at a speed V corresponding to the set magnification. As a result, the original placed on the original table glass is scanned over the entire surface. Also, the mirrors 103b, 1
Reference numeral 03c also moves in the arrow direction (sub-scanning direction) at a speed of V / 2 as the exposure lamp 101 and the mirror 103a move in the arrow direction. Multi-valued electrical signals of three colors of R, G, and B obtained by the CCD sensor 105 are processed by the read signal processing unit 106 to be yellow (Y), magenta (M),
After being converted into 8-bit gradation data of either cyan (C) or black (BK), it is output to the external output port 108 and the copying section 200. In the copying unit 200,
The image data correction unit 201 performs gradation correction (γ correction) on the input gradation data according to the gradation characteristics of the photoconductor. The printer exposure unit 202 outputs the corrected image data to D
A / A conversion is performed to generate a laser diode drive signal, and the semiconductor laser is caused to emit light by this drive signal. The laser beam generated from the printer exposure unit 202 corresponding to the gradation data exposes the photosensitive drum 204 which is rotationally driven via the reflecting mirrors 203a and 203b. Photoconductor drum 2
04 is the eraser lamp 2 before receiving the exposure for each copy.
It is irradiated at 11, and is uniformly charged by the charging charger 205. When exposed in this state, an electrostatic latent image of the original is formed on the photosensitive drum 204. Only one of the cyan (C), magenta (M), yellow (Y), and black (BK) toner developing units 206a to 206d is selected to develop the electrostatic latent image on the photosensitive drum 204. To do. The developed toner image is transferred to the copy paper wound around the transfer drum 218 by the transfer charger 209 after excess charges are removed by the pre-transfer eraser 208. A transfer film is attached to the surface of the transfer drum 218, and the transfer drum 218 rotates counterclockwise at the same speed as the rotation speed of the photoconductor. A reference plate 220a is provided inside the transfer drum 218 to synchronize the holding position of the copy sheet and the image transfer position. The reference position sensor 220b generates a predetermined reference signal every time the reference plate 220a crosses the sensor as the transfer drum 218 rotates. The copy paper is conveyed from the paper feed cassette group 212 to the conveyance path by the paper feed roller 213, and is conveyed by the conveyance roller 214.
Is conveyed to the timing roller 217. When the copy sheet is inserted from the manual feed tray 216, it is conveyed to the timing roller 217 by the conveyance roller 215. The timing roller 217 supplies the copy paper to the transfer drum 218 in synchronization with the reference signal, and transfers the copy paper to the transfer drum 2.
Hold in place on 18. Timing roller 21
The copy paper supplied from the No. 7 to the transfer drum 218 is electrostatically adsorbed to the transfer drum 218 by the adsorption charger 219. The above printing process is repeated for four colors of yellow (Y), magenta (M), cyan (C), and black (BK). At this time, the photosensitive drum 2
04, the exposure lamp 101 and the mirrors 103a, 103b, and 103c repeat predetermined operations in synchronization with the operation of the transfer drum 218. After that, the copy paper is separated from the transfer drum 218 by removing the electric charge of the paper electrostatically adsorbed by the charge removal separation charger pair 221. The copy paper separated from the transfer drum 218 has a fixing roller pair 22.
After the fixing process is performed by 3, the paper is ejected to the paper ejection tray 224.

FIG. 4 is a block diagram showing each signal processing executed by the read signal processing unit 106. Each R, G, B image data of the original read by the CCD sensor 105 has variations due to individual differences of the CCD sensor 105 included in each copying machine. Therefore, even when the reference patch of the same color table is read, the value of the read data differs for each copying machine. In the reading device color correction processing unit 601,
The read RGB image data is corrected to standard RGB image data standardized by the NTSC standard or the high definition standard. The OR, OG, and OB image data corrected by the reading device color correction processing unit 601 are
It is output to the next color space conversion processing unit 602 and also output to the external input / output port 108. The peripheral device connected to the copying machine receives the OR, OG, and OB image data of the original through the external input / output port 108. Further, in the copying machine of this embodiment, the peripheral device is connected to the external input / output port 1
It is also possible to form an image using the OR, OG, and OB image data input via 08. In this case,
The copier will function as a printer. this is,
This is because each processing unit after the reading device color correction processing unit 601 is set to use standardized RGB image data. The color space conversion processing unit 602 converts the standardized RGB image data (OR, OG, OB) into the XYZ color system and then into each data of the L * a * b * color system. FIG. 5 is a diagram showing an L * a * b * color system solid. Lightness 0 (black) to 255 (white) is represented by L *, and hue and saturation are represented by units of a * and b *. Chromaticity component a
* And b * respectively indicate the direction of color, and the chromaticity component a *
Indicates the red-green direction, and the chromaticity component b * indicates the yellow-blue direction.
Here, the reason for converting the RGB image data into the L * a * b * color system is as follows. As mentioned above, GBT
In the C method, each 8-bit image data Xij in the 4 × 4 pixel block is converted into 2-bit code data φij. When decoding, gradation range index LD and average value information LA
The 256 gradation data of four types of values specified based on and are replaced in correspondence with the code data φij assigned to each pixel. In this way, the image data obtained by decoding has a certain degree of error with respect to the image data before encoding. When the color of each pixel is reproduced by using the R, G, and B image data having these errors, the color of the edge portion of the document is shifted. However, L * a
If each data of the * b * color system is used, even if an error occurs in the value of the data to be decoded, a slight shift in brightness or chromaticity may cause a color shift at the edge of the document. Absent. Therefore, in the copying machine of the present embodiment, when encoding and decoding the image data of the document, the RGB image data is once converted into data of the L * a * b * color system. The L * a * b * color system data is used in the copying machine of this embodiment.
For the above reason, if the RGB image data is converted into the hue, lightness, and saturation data, L * u * v
* The data may be converted into data of a color system or another color system such as YCrCb or HVC. The color space optimization processing unit 603 encodes the image data at the time of decoding before encoding each data of the image information L *, a *, b * of the document represented by the L * a * b * color system. To reduce the deterioration of
The arithmetic processing based on the graphs (a) to (c) shown in FIG. 6 is executed for each data, and the distribution of the lightness component L * is changed so as to be distributed in the range of 0 to 255 for each document. , Chromaticity a * and b * component distributions of -127 to 1 for each original
The distribution is changed to 28. First, regarding each data of the image information L *, a *, b * of the manuscript,
Maximum values L * max, a * max, b * max and minimum values L * min, a * min, b * min are obtained. In the color space optimization processing unit 603, the lightness component L * is calculated by performing the operation shown in the following “Equation 5”.

[Formula 5] L1 * = 255 / (L * max-L * min)
X (L * -L * min) The calculation process is based on the graph shown in (a) of FIG. That is, in the calculation of “Equation 5” above, L * max
The value of the lightness component L * distributed in the range of
Change to a value distributed in the range of 255. Further, with respect to the chromaticity component a *, the calculation shown in the following “Equation 6” is executed to obtain the chromaticity component a1 *. In addition, in "Equation 6", when the value of a * is 0, processing is performed so that the value of a1 * after the optimization processing is also 0. This is to maintain the pixel color when it is an achromatic color in which the values of the chromaticity components a * and b * are both 0.

## EQU6 ## a1 * = 128 / a * max × a * where 0
≦ a * ≦ a * max a1 * = 127 / | a * min | × (a * -a * mi
n) -127 However, a * min ≦ a * ≦ 0 The calculation process is based on the graph shown in FIG. That is, in the calculation of “Equation 6”, 0 to a * m
Each value of a * distributed in the range of ax is changed to a value distributed in the range of 0 to 128, and each value of chromaticity component a * distributed in the range of a * min to 0 is changed to the range of -127 to 0. Change to be distributed in. Further, for the chromaticity component b *, the calculation shown in the following “Equation 7” is executed to obtain the chromaticity component b1 *.
In “Equation 7”, as in the case of “Equation 6”, when the value of b * is 0, the value of b1 * after the optimization processing is also 0. This is to maintain the pixel color when it is an achromatic color in which the values of the chromaticity components a * and b * are both 0.

## EQU7 ## b1 * = 128 / b * max × b * where 0
≦ b * ≦ b * max b1 * = 127 / | b * min | × (b * −b * mi
n) -127 However, b * min ≦ b * ≦ 0 The calculation process is based on the graph shown in FIG. That is, in the calculation of “Equation 7”, 0 to b * m
Each value of b * distributed in the range of ax is changed to a value distributed in the entire range of 0 to 128, and each value of chromaticity component b * distributed in the range of b * min to 0 is set to -127 to 0. Change so that it is distributed in the range. It should be noted that the maximum values L * max, a * max, for the respective data of the image information L *, a *, and b * of the original used in the calculations shown in the above "expression 5" to "expression 7".
b * max and minimum values L * min, a * min, b *
Each min is stored in the hard disk 614 and used when performing the color space inverse conversion process. The above arithmetic processing is performed for the following reasons. That is, GBTC
In the encoding process and the decoding process according to the method, division is frequently used, as shown in the above "Equation 1" to "Equation 4" and "Table 2". Therefore, when the difference between the component data of each pixel is small, the difference disappears during the calculation, and the reproducibility of the image data obtained by the decoding process deteriorates. In the color space optimization processing unit 603, by the above calculation,
The distribution of the lightness component L * is changed to a value distributed in the range of 0 to 255 for each document, and the distribution of each component of the chromaticity a * and b * is distributed to the range of -127 to 128 for each document. Change to a value. This alleviates the inconvenience. In the encoding / decoding processing unit 604, after executing the GBTC method encoding process, the attribute of the image is discriminated in each block unit as described in detail later, and the encoding is performed based on the discriminated result. The data (gradation width index LD, average value information LA, code data φij) is compressed again. Here, the attributes of the image refer to five of a monochrome solid image, a color solid image, a monochrome binary image, a monochrome multi-valued image, and a full color image. The compressed data obtained by the second compression is stored in the compressed image memory 610, and the flags f1, f2, and f3 indicating the attributes of the image are stored in the hard disk 614. When performing the decoding process, the CPU 611 causes the compressed image memory 610.
The compressed data corresponding to each address is read from the encoding / decoding processing unit 604 from the
The flags f1, f2, and f3 indicating the attributes of the image are read from. The encoding / decoding processing unit 604 encodes the recompressed data based on the values of the flags f1, f2, and f3 indicating the image attributes (gradation width index LD,
It is expanded to average value information LA and code data φij). Subsequently, the data encoded according to the GBTC method is decoded, and the lightness component L2 * and the chromaticity components a2 * and b2 * of the image data are output. In the color space inverse optimization processing unit 605,
The maximum values L * max, a * max, b * max and the minimum values L * min, a * min, b * min are read from the hard disk 614, and the read values are used to decrypt the decoded L2 *, a2 *. , B2 *, the original distribution of L * max to L * min, a * max to a * min, b
Return to * max to b * min. These processes are executed based on the graphs shown in (a) to (c) of FIG. 7. That is, the lightness component L2 * is subjected to the arithmetic processing of the following “Equation 8” to restore the lightness component L3 * distributed in L * max to L * min.

[Equation 8] L3 * = (L * max-L * min) / 255
For the xL2 * + L * min chromaticity component a *, the arithmetic processing of the following “Equation 9” is executed, and the chromaticity component a3 distributed in a * max to a * min.
Return to *.

A3 * = a * max / 128 × a2 * where 0 ≦ a * ≦ 128 a3 * = 127 / | a * min | × (a * + 127) +
a * min However, for -127 ≦ a * ≦ 0 chromaticity component b *, the following arithmetic processing is performed, and the chromaticity component b distributed in b * max to b * min.
Return to 3 *.

B3 * = b * max / 128 × b2 * where 0 ≦ b * ≦ 128 b3 * = 127 / | b * min | × (b * + 127) +
b * min However, in -127 ≦ b * ≦ 0, the color space inverse conversion processing unit 606 restores L3 *, a3 *, b3 in the color space inverse optimization processing unit 605.
Each data of * is inversely converted into RGB image data of OR1, OG1, and OB1. The reflection / density conversion processing unit 607
The RGB image data of R1, OG1, and OB1 are subjected to predetermined reflection / density conversion processing, and then the density data of DR, DG, and DB are output. The RGB image data converted into the density data is converted into image data of any one of cyan (C), magenta (M), yellow (Y), and black (BK) in the masking processing unit 608, and then, It is output to the image data correction unit 201. The image data correction unit 201 performs predetermined gradation correction (γ correction) processing on the gradation data output from the masking processing unit 608, and then outputs the gradation data to the printer exposure unit 202. .

(2-2) Operation Panel FIG. 8 is a front view of the operation panel 300 provided in the digital full-color copying machine of this embodiment. Operation panel 300
Is a ten-key pad 301 for inputting the number of copies, a start key 302 for starting a copying operation, a mode setting and a display section 303 for displaying the copy status, and a selection for selecting the mode displayed on the display section 303. A key 304, an enter key 305 for setting the mode displayed on the display unit 303, an image editing key 306 for selecting and setting the image editing mode, and a magnification setting key 307 for setting the copy magnification.
And a paper selection key 308 for selecting the size of the copy paper,
An automatic setting key 309 for setting the automatic paper selection function is provided. In the copying machine of this embodiment, the editing / processing process to be executed can be set according to the attribute of the image determined in the unit of 4 × 4 pixel blocks. The contents of the edit / modification processing are displayed on the display unit 303 in response to the pressing of the image edit key 306.
Determined by the settings of the items displayed in. Selection and setting of items displayed on the display unit 303 are performed as follows. First, each item displayed on the display unit 303 is selected by the selection key 304 and highlighted. Next, pressing the enter key 305 completes the setting of the selected item. Along with the setting of items for editing / processing, the mode flag MF of the set items is set as shown in "Table 3" below.
Is set to "1".

[Table 3] For example, by inverting the "mono color" item with the selection key 304 and pressing the enter key 305, mono color conversion processing is set, and the overall flag MF1 and the mode flag MF51 for the full color image are set to "1". To be done. In this case, in the editing / processing process described later, all the gradation width indexes LD of the chromaticity components a * and b *, the average value information LA, and the code data value for the full-color image are replaced with 0. For the mono-color conversion process, color conversion process, erase process, negative / positive inversion process, background cut process, and dpi conversion process, data encoded by the GBTC method (gradation width index LD, average value information LA,
The code data φij) is used. Therefore, it is possible to reduce the memory required for the editing process as compared with the conventional case. The sharpness processing, the gradation curve correction processing, and the outline trimming processing are performed after the RGB image data is decoded as usual.

(3) Description of Image Processing (3-1) Main Routine FIG. 9 is a main routine of copy processing executed by the CPU 611 of the copying machine of this embodiment. First, the main body of the copying machine is initialized (step S1000). Next, key input processing from the operation panel 300 is performed (step S200).
0). Next, warm-up of the apparatus, shading, pre-processing such as image stabilization processing is executed (step S
3000). After that, the CPU 611 drives the scanner motor 102 to read the image data of the original placed on the original table 107, standardizes the RGB image data obtained by the reading, and then sets the standardized RGB image data to L The data is converted into * a * b * color system data (step S4000). The original image data represented by the L * a * b * color system is subjected to the encoding process using the GBTC method, and then stored in the compressed image memory 610 (step S5000). In the copying machine of this embodiment, the encoded data (gradation width index L) stored in the compressed image memory 610 is used.
D, average value information LA, code data φij),
The attribute of the image of the 4 × 4 pixel block to which the encoded data belongs is discriminated, and the encoded data is further compressed based on the discrimination result (step S6000).
Here, the image attributes are 5 of a monochrome solid image, a color solid image, a monochrome binary image, a monochrome multivalued image, and a full color image.
Say one. A flag f indicating the attribute of each determined block
1, f2 and f3 are stored in the hard disk 614. The recompressed data is stored in the compressed image memory 6
Stored in 10. In the next step S7000, the data stored in the compressed image memory 610 is read and the flags f1, f2 and f indicating the attribute of the data are read out.
3 is read from the hard disk 614, and the recompressed data is the original encoded data (average value information LA, gradation range exponent LD, code data φij) based on the image attribute determined from the value of each flag. Extend to. AE processing for optimizing the density distribution of an image, and editing / processing processing such as color conversion and frame editing based on the gradation width index LD of encoded data obtained by expansion and the average value information LA To do. Encoded data (average value information LA, gradation range index LD,
The coded data φij) is stored in the compressed image memory 610. In the next step S8000, the compressed image memory 61
The RGB image data is read by reading the coded data from 0, decoding the data into 256 gradation data for the lightness component L *, the chromaticity components a * and b *, and further executing the color space inverse conversion process. Return to. In step S9000, an image forming process for forming an image on a sheet is executed based on the RGB image data obtained by the decoding process.
After the image forming process, a process necessary for maintaining the condition of the apparatus, such as removal of the residual toner on the photosensitive drum 204 after the image formation, which is not directly related to the image forming operation, is performed (step S9800). Finally, although not directly related to the image forming process of this embodiment, temperature control and communication control of the fixing device are performed (step S9900).

(3-2) Key Input Processing FIG. 10 shows the key input processing (step S200 shown in FIG. 9).
It is a processing flowchart of 0). When a key input is made on the operation panel 300 (step S20)
If YES at 01), the following processing is executed according to the type of the input key. The input key is the image edit key 306
If (YES in step S2002), the display unit 303
The image editing menu screen shown in FIG. 8 is displayed (step S2003). The selection of the 11 edit items displayed is
This is done by operating the cursor key 304. The selected item is highlighted in black and white. On the screen shown in FIG. 8, the item "mono color" is selected. Enter key 305
By pressing, the selected item is set.
The set item maintains the black-and-white inverted state even when another item is selected by operating the cursor key 304. When the mode setting is performed by the user (YES in step S2004), flag setting processing is executed (step S2005). In the flag setting process,
The mode flag MF of the set item is set to "1". The mode is not set (N in step S2004
O), if the end item is selected (step S200)
After the screen of the display unit 303 is returned to the initial screen (step S2007), key input is again waited (step S2001). If the input key is the print key 302 (YES in step S2008)
S), a predetermined image forming start process is executed (step S20).
09), and return. When the input key is neither the image edit key 306 nor the print key 302 (NO in step S2008), other processing is executed (step S2010), and waiting for key input again (step S2010). S2001).

The above "Table 3" shows the items selected by the user and the types of flags set to "1" by the flag setting process (step S2005) corresponding to the set items. The mode flags are all set to "0" in the initial setting (step S1000). For example, when the user sets the item "mono-color conversion", the overall flag MF1 meaning that the item is set and the mode flag MF51 of the full-color image to which the conversion process is applied are "1". Is set to. When the item of “color conversion” is set, the whole flag MF2 meaning that the item is set, and the mode flag MF32 for the monochrome binary image and the full-color image to which the conversion process is applied are set. And the MF52 is set to "1". When the "erase" item is set, the overall flag MF3 indicating that the item is selected, and the black and white solid image, the color solid image, the black and white binary image, and the black and white multi-valued image to which the processing is applied , Mode flags MF13, MF23, MF33, MF43 and MF53 for full color images are set to "1". Hereinafter, each item is as shown in "Table 3".

(3-3) Recompression Process FIG. 11 is a flowchart relating to the recompression process of data encoded by the GBTC method (step S6000 shown in FIG. 9). Here, the attribute of the image is determined from the average value information LA of the data compressed and encoded by the GBTC method and the value of the gradation width index LD. Next, recompression processing is performed to reduce the data that can always be specified based on the image attribute among the lightness component L, the chromaticity components a * and b *, and the code data φij. As a result, the compression rate of image data can be improved without sacrificing reproducibility. Even when the reproducibility is emphasized and the number of bits of the code data φij assigned to each pixel is increased, the reduction of the compression rate can be prevented by executing the recompression process. First, the code data stored in the compressed image data memory 610 is read (step S600).
1), attribute determination processing is executed (step S600)
2). In this attribute discrimination processing, for each code data, which of the monochrome solid image, the color solid image, the monochrome binary image, the monochrome multi-valued image, and the full-color image the 4 × 4 pixel block relating to the code data belongs to? And a flag f1 for a solid image, a flag f2 for a color / monochrome image, and a flag f for a binary / multivalued image.
Set the values of 3 respectively. The attribute determination process will be described later with reference to a flowchart. Next step S6
In 003, the value of each flag is checked, and the following processing is executed according to the type of each image. 4 × 4 pixel block is a black and white solid image (flag f1 = 1, flag f2 = 1)
If it belongs to, the black and white solid image feature amount extraction processing (step S6004) is executed. Here, the histogram data of the average value information LA of the lightness component L * is created. next,
A black and white solid image recompression process is executed (step S600).
5). When the 4 × 4 pixel block belongs to the color solid image (flag f1 = 1, flag f2 = 0), the color solid image feature amount extraction processing (step S6006) is executed. Here, each histogram data of each average value information LA of the lightness component L * and the chromaticity components a * and b * is created. Next, a color solid image recompression process is executed (step S6007). The 4 × 4 pixel block is a monochrome binary image (flag f1 = 0, flag f2 = 1, flag f
3 = 1), a monochrome binary image feature amount extraction process (step S6008) is executed. Here, 4x4
The black and white ratio of each pixel existing in the pixel block is calculated. Next, a monochrome binary image recompression process is executed (step S6).
009) 4 × 4 pixel block is a black and white multi-valued image (flag f1 =
0, flag f2 = 1, flag f3 = 0)
A black-and-white multi-valued image feature amount extraction process (step S6010) is executed. Here, the histogram data of the average value information LA of the lightness component L * and the gradation width information LD is created. After that, the black-and-white multi-value image recompression process is executed (step S
6011). A 4 × 4 pixel block is a full-color image (flag f1 = 0, flag f2 = 0, flag f3 = 1).
If it belongs to, the full-color image feature amount extraction processing (step S6012) is executed. Here, the lightness component L *,
The histogram data of the average value information LA of the chromaticity components a * and b * and the gradation width index LD is formed. 4x4 above
After performing the process according to the type of image to which the pixel block belongs, the compressed data obtained by the recompression process is stored in the compressed image memory 610 (step S6013). However,
The encoded data regarding the full-color image is stored in the compression code memory 611 as it is. The above processing (steps S6001 to S6013) is executed for all the coded data of 4 × 4 pixel blocks of the original. When the above process is performed on the encoded data of all 4 × 4 pixel blocks (step S6
(YES at 014), the process is terminated and the process returns.

(3-3-1) Attribute Discrimination Processing FIG. 12 shows the attribute discrimination processing (step S60 shown in FIG. 11).
02) is a flow chart. First, it is determined whether or not the 4 × 4 pixel block corresponding to the encoded data belongs to the solid image (step S6020). In the solid image discrimination processing, when it is determined that the 4 × 4 pixel block belongs to the solid image, the value of the flag f1 is set to 1, and when it is determined that the block does not belong to the solid image, the value of the flag f1 is set to 0.
Set to. After the completion of the solid image discrimination processing, the value of the flag f1 is checked (step S6021). Here, the flag f
When the value of 1 is 1, it is determined that the 4 × 4 pixel block corresponding to the code data belongs to the solid image, and the following processing is executed. First, color / monochrome discrimination processing is executed (step S6022). In the color / monochrome discrimination processing, when it is determined that the 4 × 4 pixel block belongs to the monochrome image, the value of the flag f2 is set to 1, and when it is determined that the block belongs to the color image, the value of the flag f2 is set to 0. .
After the color / monochrome image discrimination processing is completed, the value of the flag f2 is checked (step S6023). Here, the flag f
If the value of 2 is 1, a process of writing an attribute signal representing a monochrome solid image in the hard disk 614 in association with the code data is performed (step S6024). Also, the flag f
When the value of 2 is 0, the attribute signal representing the color solid image is associated with the code data and written in the hard disk 614 (step S6025). In step S6020, if the value of the flag f1 is 0, it is determined that the 4 × 4 pixel block belongs to an image having two or more gradations (NO in step S6021), and the image is color / monochrome. Which image of
In the case of a monochrome image, it is checked whether this is a binary image or a multi-valued image. First, color / monochrome discrimination processing is executed (step S6026). In the color / monochrome discrimination processing, when it is determined that the 4 × 4 pixel block belongs to the monochrome image, the value of the flag f2 is set to 1, and when it is determined that the block belongs to the color image, the value of the flag f2 is set to 0. Set. The color / monochrome discrimination processing in step S6026 is the same as in step S6022 above.
It is exactly the same as the processing in. After the color / monochrome image discrimination processing is completed, the value of the flag f2 is checked (step S6027). Here, when the value of the flag f2 is 1, a determination process is continuously executed as to whether the monochrome image is a binary image or a multi-valued image (step S).
6028). In the binary / multi-value discrimination processing, when it is determined that the 4 × 4 pixel block belongs to the binary image, the value of the flag f3 is set to 1, and when it is determined that the block belongs to the multi-valued image, the value of the flag f3 is changed. Set to 0. After the binary / multivalue discrimination processing is completed, the value of the flag f3 is checked (step S6029).
Here, when the value of the flag f3 is 1, a process of writing the attribute signal representing the monochrome binary image in the hard disk 614 in association with the encoded data is performed (step S6).
030). If the value of the flag f3 is 0, the attribute signal representing the monochrome multi-valued image is written in the hard disk 614 in association with the coded data (step S6031). When the value of the flag f2 set in the color / monochrome discrimination processing in the above step S6026 is 0, it is determined that the image is a full-color image (NO in step S6027). Therefore, the attribute signal representing the full-color image is written in the hard disk 614 in association with the encoded data (step S60).
32). After the above process is executed for all the encoded data stored in the compressed image memory 610 (YES in step S6033), the process returns.

<3-3-1-1> Solid Image Discrimination Processing FIG. 13 shows a solid image discrimination processing (step S shown in FIG. 12).
6020) is a flowchart. The solid image is an image composed of only one color having a certain lightness component L * and chromaticity components a * and b *. Visually, the lightness L
When the value of each gradation width index LD of *, chromaticity a * and b * is less than or equal to a predetermined value, it is recognized as a solid image. However, when the value of the average value information LA of the lightness component L * is high to some extent (for example, 240 or more), the value of the gradation width index LD of the lightness L * and the chromaticities a * and b * is larger than the above predetermined value. Even a value is recognized as a solid image. In the solid image discrimination processing of the present embodiment, the value of the average value information LA of the lightness component L * is less than 240, and the lightness L * and the chromaticity a are
When the value of the gradation range index LD of each component of * and b * is 2 or less (all YE in steps S6040 to S6043)
S), or the value of the average value information LA of the lightness component L * is 24
0 or more (NO in step S6040), and lightness L
When the value of the gradation width index LD of each component of *, chromaticity a * and b * is 6 or less (YES in steps S6046 to S6048), it is determined that the 4 × 4 pixel block belongs to the solid image, and the flag f1. Is set to 1 (step S60
44, or S6049). Lightness component L *, Chromaticity component a
When the value of each gradation width index LD of * and b * does not satisfy the above condition (the value of the average value information LA of the lightness component L * is 240
Is less than (YES in step S6040), and even if any one of steps S6041 to S6043 is NO, or the value of the average value information LA of the lightness component L * is 240 or more (step S6040). NO), any one of steps S6046 to S6048 is N
When there is O), it is determined that the 4 × 4 pixel block does not belong to the solid image, and the value of the flag f1 is set to 0 (step S6045 or S6050).

<3-3-1-2> Color / Monochrome Discrimination Processing FIG. 14 is a flowchart showing the color / monochrome discrimination processing (steps S6022 and 6026 shown in FIG. 12). As illustrated, the processing executed in step S6022 and the processing content executed in step S6026 are the same. Here, the color / monochrome discrimination processing in step S6022 will be sequentially described. The black and white image is an achromatic color in which the data values of the chromaticity components a * and b * are almost zero. For example, when the value of the average value information LA of each of the chromaticity a * and b * components is within ± 5, it can be determined that the image has an achromatic color. However, with only the above-mentioned criteria, the average value information LA of the chromaticity components a * and b * is within ± 5, and the chromaticity component a of a pixel is
When the data value of * is -120 and the data value of the chromaticity component a * of other pixels is 125 and the average value information LA has a small value, this is erroneously determined to be a monochrome image. Therefore, the CPU 611 of the present embodiment sets the gradation width index L
Only when the value of D is within ± 5, it is determined that the 4 × 4 pixel block belongs to the monochrome image. Specifically, each average value information LA of the chromaticity components a * and b * and gradation width information LD
Are both within ± 5 (steps S6051 to S6051
If YES in S6054, the value of the flag f2 is set to 1 (step S6055). If even one of the average value information LA of the chromaticity components a * and b * and the gradation width information LD is ± 5 or more (NO in any one of steps S6051 to 6054), 4 × 4 It is determined that the pixel block belongs to the color image, and the value of the flag f2 is set to 0 (step S6056).

<3-3-1-3> Binary / Multivalue Discrimination Processing FIG. 15 is a flowchart showing the binary / multivalue discrimination processing (step S6028 shown in FIG. 12). In the encoded data of the binary image 4 × 4 pixel block, the value of the gradation width index LD of the lightness component L * is large, and the code data φij of the lightness component L * is 10 or 00 representing halftone data. not exist. On the other hand, 4 × 4 belonging to a multi-valued image
In the code data φij of the lightness component L * of the pixel block,
There are 10 or 00 representing halftone data. Therefore, in the binary / multi-value discrimination process, the value of the gradation width index LD of the lightness component L * is 200 or more (step S607).
If YES, the value of all code data φij of the lightness component L * is neither 10 nor 00 (step S60).
71 and NO in S6072, and YE in S6073
Only in S), it is determined that the 4 × 4 pixel block belongs to the binary image, and the value of the flag f3 is set to 1 (step S6).
074). On the other hand, the value of the gradation width index LD of the lightness component L * is 200 or less (NO in step S6070),
For the value of the code data φij of the lightness component L *, even one is 10
Or if there is 00 (steps S6071, S6
If either one of 072 is YES), it is determined that the 4 × 4 pixel block belongs to the multi-valued image, and the value of the flag f3 is set to 0 (step S6075).

(3-3-2) Feature Amount Extraction Process The average value information LA and the gradation range index LD obtained for each block by the GBTC system encoding process are the information amount of 1/16 of the original image data. , Average value information in all image data, and a representative value of gradation width data. Therefore, the lightness component L * and the chromaticity components a * and b of the original image data are actually
About the same information as when the feature quantity of * is calculated is 1 /
It will be obtained from 16 data. This makes it possible to simplify the arithmetic circuit required for the feature amount extraction processing and reduce the time required for the feature amount extraction processing. In the copying machine of the present embodiment, the feature amount extraction processing described below is executed, and each extracted feature amount is stored in the hard disk 614. This makes it possible to read the corresponding feature amount from the hard disk and execute each process based on the read value when performing the AE process or the image editing / modifying process described later. Therefore, it is possible to reduce the memory capacity required for the processing as compared with the case where the AE processing or the image editing / modification processing is executed based on the image data before being encoded and after being decoded. Further, the time required for the processing itself can be shortened.

<3-3-2-1> Monochrome solid image feature amount extraction processing FIG. 16 is a flowchart showing the monochrome solid image feature amount extraction processing (step S6004 shown in FIG. 11). Here, after the histogram data of the average value information LA for all the blocks of the lightness component L * is created (step S6100), the process returns. The histogram data obtained here is stored in the hard disk 614 and is used when executing the background cutting process which is one of the editing / processing processes described later.

<3-3-2-2> Color Solid Image Feature Amount Extracting Process FIG. 17 is a flowchart showing the color solid image feature amount extracting process (step S6006 shown in FIG. 11). First, histogram data of average value information LA for all blocks of the lightness component L * is created (step S
6110). Next, the histogram data of the average value information LA for all the blocks of the chromaticity component a * is created (step S6111). Next, the histogram data of the average value information LA for all the blocks of the chromaticity component b * is created (step S6112). Each of the histogram data obtained here is stored in the hard disk 614 and is used when executing the background cutting process which is one of the editing / processing processes described later.

<3-3-2-3> Monochrome Binary Image Feature Amount Extraction Process FIG. 18 is a flowchart showing the monochrome binary image feature amount extraction process (step S6008 shown in FIG. 11). Here, the black / white ratio of each pixel existing in the 4 × 4 pixel block is obtained. When the value of the higher 1 bit of the code data of each pixel in the 4 × 4 pixel block is 1 (YES in step S6120), the pixel is determined to be white, and the white pixel is counted up (step S61).
21). If the value of the upper 1 bit of the code data is 0 (NO in step S6120), the pixel is determined to be black and the black pixel is counted up (step S6122). The judgment in step S6120 is that the code data φij = 11,10 is assigned when the value of the 256 gradation data is larger than the value of the average value information LA, and φij = 00,01 is assigned when the value is small. based on. After making a judgment for all the code data in the 4 × 4 pixel block (step S61)
If YES in 23, the black-and-white ratio is calculated (step S61).
24). Here, the obtained black-and-white ratio is stored in the hard disk 614.

<3-3-2-4> Monochrome multi-valued image feature amount extraction process FIG. 19 is a flowchart showing the black-and-white multi-valued image feature amount extraction process (step S6010 shown in FIG. 11). First, histogram data of average value information LA for all blocks of the lightness component L * is created (step S
6130). Next, the histogram data of the gradation width information LD for all the blocks of the lightness component L * is created (step S6131). The respective histogram data obtained here are stored in the hard disk 614 and used in the AE processing described later.

<3-3-2-5> Full Color Image Feature Extraction Process FIG. 20 is a flowchart showing the full color image feature extraction process (step S6012 shown in FIG. 11). First, histogram data of average value information LA for all blocks of the lightness component L * is formed (step S
6140). Next, histogram data of the gradation width index LD for all blocks of the lightness component L * is formed (step S6141). Histogram data of the average value information LA for all blocks of the chromaticity component a * is formed (step S6142). Histogram data of the gradation width index LD for all blocks of the chromaticity component a * is formed (step S6143). The histogram data of the average value information LA for all the blocks of the chromaticity component b * is formed (step S6144). Histogram data of the gradation width index LD for all blocks of the chromaticity component b * is formed (step S6145). The respective histogram data obtained here are stored in the hard disk 614 and used in the AE processing described later.

(3-3-3) Recompression Processing In the copying machine of the present embodiment, the average value information LA about the lightness component L *, the chromaticity components a * and b *, the gradation width index LD and the code data φij. Of the above, the attribute determination process (step S
Based on the attribute information determined in 6002), the data that can be always specified is reduced, and the image data is further compressed. The data obtained by the recompression process is stored in the compressed image memory 610. The recompression process executed for each attribute will be described below.

<3-3-3-1> Monochrome Solid Image Recompression Processing FIG. 21 is a processing flowchart of the monochrome solid image recompression processing (step S6005 shown in FIG. 11). The black and white solid image is an achromatic image having a specific lightness component L * (the data values of the chromaticity components a * and b * are both 0). That is, a black and white solid image can be reproduced from the value of the specific lightness component L *. Since the value of the lightness component L * is a specific value, the values of the code data φij assigned to each pixel in the 4 × 4 pixel block are all the same code.
Further, the value of the gradation width index LD becomes zero. Therefore, only the average value information LA of the lightness component L * is taken out from the compressed image memory 610, and this is stored in the compressed image memory 610 as recompressed data for the black and white solid image (step S).
6200). FIG. 22 is a conceptual diagram showing the relationship between encoded data and recompressed data. The recompressed data is compressed to 1/18 of the coded data before recompression. Since the compression rate of the encoded data obtained by the GBTC method is 3/8, the compression rate of the recompressed data is 1/48.

<3-3-3-2> Color Solid Image Recompression Processing FIG. 23 is a processing flowchart of the color solid image recompression processing (step S6007 shown in FIG. 11). A color solid image is an image of a single color (the data values of chromaticity components a * and b * are constant) having a specific lightness component L *. That is, the color solid image can be reproduced from the average value information LA of the chromaticity components a * and b * and the average value information LA of the lightness component L *. Therefore, the compressed image memory 61
Only the average value information LA of the lightness component L * and the chromaticity components a * and b * is extracted from 0 and stored in the compressed image memory 610 as recompressed data regarding the color solid image (steps S6210, S6211, S6212). ).
FIG. 24 is a conceptual diagram showing the relationship between coded data and recompressed data. The recompressed data is compressed to 1/6 of the coded data before recompression. Since the compression rate of encoded data obtained by the GBTC method is 3/8, the recompressed data is 1/16.

<3-3-3-3> Monochrome Binary Image Recompression Process FIG. 25 is a flowchart of the monochrome binary image recompression process (step S6009 shown in FIG. 11). The black-and-white binary image is an achromatic image (both the data values of the chromaticity components a * and b * are 0) consisting of black (the lightness component L * is 0) or white (the lightness component L * is 255). . That is, black and white 2
The value image can be reproduced from the code data φij for the lightness component L *. In the case of a monochrome binary image, the code data φij assigned to each pixel is composed of only two codes 11 or 01. Therefore, the value of the code data φij for the lightness component L * is fetched from the compressed image memory 610, the code data of each pixel thus fetched is converted into 1-bit data φ′ij, and the converted 1-bit data is regenerated. Use as compressed data. The value of the 1-bit data φ′ij is, for example, 1 when φij = 11 and 0 when φij = 01. As the processing flow, first, one block of 6-byte data of the lightness component L * is read from the compressed image memory 610, and the 2-byte data of the average value information LA and the gradation width information LD is removed from the read data, Only 4-byte data indicating the code data assigned to each pixel is set (step S622).
0). Next, the leading 2-bit data of the 4-byte data is read, and this data is set as code data φij (step S6221). In the case of a binary image, 11 or 01 code data φij is assigned to each pixel. Therefore, the value of the read 2-bit data φij is checked (step S6222). When φij = 01, φ′ij = 0 is assigned (step S622).
3). If φij = 11, φ′ij = 1 is assigned (step S6224). Next, the 2-bit data at the head of the 4-byte data is shifted, and the 2-bit data next to the 2-bit data is set at the head (step S6225). Step S6221 above
The processing from S6225 is performed for all 16 pixels. As a result, recompressed data φ′ij is obtained. The recompressed data φ′ij is stored in the compressed image memory 610.
When the first 2-bit data is at the beginning again by the shift operation in step S6225 (step S62
If YES at step 26), the process is terminated and the process returns. FIG. 26
FIG. 3 is a conceptual diagram showing the relationship between coded data and recompressed data. The recompressed data consists of 1-bit code data φ′ij × 16 pixels = 16-bit data,
It is compressed to 1/9 of the code data before recompression. GBTC
Since the compression rate of the code data obtained by the method is 3/8, the compression rate of the recompressed data is 1/24.

<3-3-3-4> Monochrome Multi-valued Image Recompression Processing FIG. 27 is a flowchart of the black-and-white multivalued image recompression processing (step S6011 shown in FIG. 11). Achromatic colors (both chromaticity components a * and b * are 0) in black-and-white multi-valued images
That is, it is an image in which the lightness component L * changes in the range of 0 to 255. That is, the black-and-white multivalued image has a lightness component L
It can be reproduced only with *. Therefore, from the compressed image memory 610, all information of the lightness component L * (average value information LA,
The gradation width index LD and the code data φij) are taken out and stored in the compressed image memory 610 as recompressed image data (step S6230). FIG. 28 is a conceptual diagram showing the relationship between coded data and recompressed data. The recompressed data is compressed to 1/3 of the code data before recompression. The compression rate of code data obtained by the GBTC method is 3
Since it is / 8, the compression rate of the recompressed data is ⅛.

(3-4) Decompression process from recompression The recompressed data is stored in the compressed image memory 610. When forming an image on a sheet, the CPU 611 reads the recompressed data stored in the compressed image memory 610, and also reads the attribute information related to the data from the hard disk 614. Decompression processing of the recompressed data is executed based on the attribute of the recompressed data specified by the read attribute information.

<3-4-1> Decompression Process from Recompression FIGS. 29 and 30 are flowcharts of the recompression and decompression process (step S7000 shown in FIG. 9). First, the attribute flags f1, f2, and f3 for the 4 × 4 pixel block to be expanded are read from the hard disk 614 (step S7001). Next its attribute flag f1,
From the values of f2 and f3, the data regarding the 4 × 4 pixel block is a monochrome solid image, a color solid image, a monochrome binary image
It is specified whether the image belongs to the black-and-white multi-valued image or the full-color image (step S7002). 4x4
When the data related to the pixel block is data that belongs to a black and white solid image (solid image flag f1 = 1, color /
Monochrome image flag f2 = 1), decompression processing is performed from the black and white solid image recompressed data, and data that can be decoded by the GBTC method is created (step S7003).
After that, the editing / modifying process based on the setting made by the user is executed (step S7004). After the above processing, the decompressed data is written in the compressed image memory 610 (step S7015). When the data related to the 4 × 4 pixel block belongs to the color solid image (solid image flag f1 = 1, color / monochrome image flag f2 =
0), decompression processing is performed from the color solid image recompressed data to create data that can be decoded by the GBTC method (step S7005). After that, the editing / processing process based on the setting made by the user is executed (step S70).
06). After the above processing, the decompressed data is written in the compressed image memory 610 (step S7015). 4x4
When the data relating to the pixel block is data belonging to a monochrome binary image (solid image flag f1 = 0, color / monochrome image flag f2 = 1, binary / multivalue image flag f3
= 1), decompression processing is performed from the black and white binary image recompressed data to create data that can be decoded by the GBTC method (step S7007). After that, the editing / processing process is executed based on the setting made by the user (step S7).
008). After the above processing, the decompressed data is written in the compressed image memory 610 (step S7015). 4x
When data relating to a 4-pixel block belongs to a black-and-white multivalued image (solid image flag f1 = 0, color /
Monochrome image flag f2 = 1, binary / multivalued image flag f
3 = 0), decompression processing is performed from the black-and-white multivalued image recompressed data to create data that can be decoded by the GBTC method (step S7009). The decompressed data is subjected to the AE processing corresponding to the black-and-white multivalued image (step S7010). After that, the editing / modifying process based on the setting made by the user is executed (step S7011). After the above processing, the decompressed data is written in the compressed image memory 610 (step S7015). When the data related to the 4 × 4 pixel block is data that belongs to a full-color image (solid image flag f1 = 0, color / monochrome image flag f2 = 0), decompression processing is performed from the full-color image recompressed data, and the GBTC method is used. Data that can be decrypted by is created (step S7012). AE processing corresponding to a full-color image is performed on the expanded data (step S7013). After that, the editing / processing process based on the setting made by the user is executed (step S7).
014). After the above processing, the decompressed data is written in the compressed image memory 610 (step S7015).

<3-4-1-1> Decompression processing from black and white solid image recompression FIG.
10 shows a flowchart of step S7003) shown in FIG.
In the case of a black and white solid image, the data stored in the compressed image memory 610 is only 1-byte (8-bit) data which is the average value information LA of the lightness component L *, as shown in the upper part of FIG. First, the recompressed data is read from the compressed image memory 610, and this is designated as data A (step S
7100). Data A is the average value information L of the lightness component L *
A (step S7101). A black and white solid image is an achromatic image. Therefore, the data of the chromaticity components a * and b * are all 0. Based on the characteristics of the black and white solid image, the average value information LA of the chromaticity components a * and b *, the gradation width index LD, and the value of the code data φij are all set to 0 (steps S7102 and S7103). The lower part of FIG. 32 shows the code data obtained in steps S7100 to S7103.

<3-4-1-2> Decompression Processing from Color Solid Image Recompression Data FIG. 33 is a flowchart of decompression processing from color solid image recompression data (step S7005 shown in FIG. 29). For color solid images, compressed image memory 610
34, the average value information LA (8 bits) of each of the lightness component L * and the chromaticity components a * and b * is stored in order in FIG. First, the first 1 byte of the recompressed data is read from the compressed image memory 610, and this is designated as data A (step S7110). The data A is set as the average value information LA of the lightness component L * (step S7111).
0 is assigned to the gradation width index LD of the lightness component L * and the code data φij (step S7112). Next, the next 1 byte of the recompressed data is read, and this is set as data B (step S7113). The data B is set as the average value information LA of the chromaticity component a * (step S7114). 0 is assigned to the gradation width index LD of the chromaticity component a * and the code data φij (step S7115). Next, the last 1 byte of the recompressed data is read and this is set as data C (step S7116). The data C is set as the average value information LA of the chromaticity component b * (step S7117). 0 is assigned to all the gradation width index LD of the chromaticity component b * and the code data φij (step S7118). The lower part of FIG. 34 shows the code data obtained in steps S710 to S7118.

<3-4-1-3> Decompression Processing from Re-compressed Monochrome Binary Image FIG. 30 is a flowchart of the decompression processing from recompressed monochrome binary image data (step S7007 shown in FIG. 29). The black and white binary image is composed of achromatic pixels whose lightness component L * is 255 or 0. Since the color of each pixel is binary black and white, the chromaticity components a * and b * are both 0.
The recompressed data φ′ij shown in the upper part of FIG. 37 is 2-byte data regarding the lightness component L *, and each 1-bit data forming the data is the color of each pixel in the 4 × 4 pixel block. Indicates white (lightness component L * is 255) or black (lightness component L * is 0). The expansion from the recompressed data to the coded data is performed by the following procedure. First, 1-byte data representing "127" in decimal is given to the average value information LA of the lightness component L * (step S7120). 10 in the gradation width information LD of the lightness component L *
1-byte data indicating "255" in a base number is given (step S7121). The 2-byte data stored in the compressed image memory 610 is read, and the read data is set as the data A (step S7122). Next, the leading 1-bit data of the data A is read out and set as the data B (step S7123). The value of data B is checked (step S7124). When the value of the data B is 0, φij of the corresponding lightness component L * is set to 01 (step S7125). Further, when the value of the data B is 1, the corresponding φij is set to 11 (step S
7126). The first 1-bit data of the data A is shifted to the end, and the next 1-bit data is set to the first (step S7127). The processing of steps S7123-S7127 is executed 16 times, that is, for all 2-byte data (step S7128). Since the pixels in the block are achromatic, the average value information LA of the chromaticity components a * and b *, the gradation width index LD, and the code data are all set to 0 (steps S7129 and S7130). Fig. 36
The lower part shows code data obtained by expanding the recompressed data stored in the compressed image memory 610.

<3-4-1-4> Decompression Processing from Black-and-white Multi-valued Image Recompression FIG. 37 is a flowchart showing the decompression processing from the black-and-white multi-valued image re-compressed data (step S7009 shown in FIG. 30). Is. A monochrome multi-valued image is composed of achromatic pixels. The recompressed data of the black-and-white multivalued image consists of a total of 6 bytes of data, as shown in the upper part of FIG. Specifically, the average value information LA (1 byte) of the lightness component L *,
Gradation width index LD (1 byte) and code data φij (4
Bytes). First, 1 from the compressed image memory 610
Data of bytes is read and this is set as data A (step S7140). The data A is set as the average value information LA of the lightness component L * (step S7141). The next 1-byte data is read from the compressed image memory 610 and designated as data B (step S7142). Data B
Is the gradation width information LD of the lightness component L * (step S7).
143). The remaining 4 bytes of data are read from the compressed image memory 610 and set as data C (step S
7144). The data C is the code data φi of the lightness component L *
j (step S7145). Since each pixel of a black-and-white multi-valued image is an achromatic color, the average value data LA (1 byte) of the chromaticity components a * and b *, the gradation width index LD (1 byte), and the code data φij (4 bytes) are , All are set to 0 (steps S7146, S7147). The lower part of FIG. 38 shows code data obtained by expanding the recompressed data stored in the compressed image memory 610.

<3-4-1-5> Decompression processing from full-color image recompression When the data regarding the 4 × 4 pixel block belongs to the full-color image (solid image flag f1 = 0,
The color / monochrome image flag f2 = 0), and the compressed image memory 610 stores 6-byte data of the lightness component L * and the chromaticity components a * and b * as they are. Therefore, for full-color images, the compressed image memory 610
Data of 6 bytes in total of 1 byte, 1 byte, and 4 bytes are sequentially read as one set, the first 1-byte data of the first set is used as the average value information LA of the lightness component L *, and the next 1-byte data is the lightness. Let the gradation width index LD of the component L *, and the next 4-byte data be the code data φij assigned to each pixel. Also, the first 1-byte data of the second set is the average value information LA of the chromaticity component a *, the next 1-byte data is the gradation width index LD of the chromaticity component a *, and the next 4-byte data is Of code data φi assigned to each pixel
j. Also, the first read 1-byte data of the third set is set as the average value information LA of the chromaticity component b *, and the next 1
The byte data is the gradation width index LD of the chromaticity component b *,
The next 4-byte data is set as code data φij assigned to each pixel.

(3-4-2) AE Processing <3-4-2-1> Monochrome Multi-Valued Image AE Processing FIG. 39 is a processing flowchart of AE processing (step S7010 shown in FIG. 30) for a black-and-white multi-valued image. Is. Here, the value of the average value information LA of each block is converted into an appropriate value based on the histogram of the average value information LA of the lightness components of all pixel blocks of the original document, and the density distribution of the image is optimized. First, the histogram data of the average value information LA of the lightness component L * is read from the hard disk 614, and the minimum value M of the black and white multivalued image portion is read from the read data.
in, the maximum value Max and the average value Ave are obtained (step S7200). Here, the minimum value Min is a density value when the frequencies are sequentially accumulated from the high density side and the accumulated value exceeds 2% of the total frequency. The maximum value Max is obtained by accumulating the frequencies in order from the high density side, and the accumulated value is 9 of the total frequencies.
The density value when it exceeds 8%. Thereby, the irregular data is removed. The average value Ave is the sum of the values obtained by multiplying the frequency of each concentration value by
It shall be divided by the total frequency. Next, the minimum value Min,
The following AE processing is executed using the maximum value Max and the average value Ave (step S7201). The AE process is executed based on the graph shown in FIG.
The value of the average value information LA distributed in the range of the maximum value Max is set to 0.
Change so that it is distributed in the range of up to 255. Specifically, the arithmetic processing shown in the following "Equation 11" is performed on the average value information LA. Average value information L obtained by calculation
Replace A ′ with the value of the original average value information LA.

LA ′ = 255 / (Max-Min) × (L
A-Min) Further, the graph of FIG. 41 shows that the average value information LA when executing another AE process and the average value information L after the AE process.
The relationship with A "is shown. AE based on the graph shown in FIG.
In the processing, the value of the minimum value Min is corrected to 0, the value of the average value Ave is corrected to 128, and the value of the maximum value Max is corrected to 255. Specifically, the arithmetic processing shown in the following "Equation 12" is performed on the average value information LA. The average value information LA ″ obtained as a result of the calculation is replaced with the value of the original average value information LA.

[Equation 12] LA ″ = 128 / (Ave-Min) × (L
A-Min) However, Min ≦ LA ≦ Ave LA ”= 127 / (Max-Ave) × (LA-Av
e) +128 However, according to the AE processing, Ave ≦ LA ≦ Max, the reproducibility of halftone can be further improved as compared with the AE processing executed based on the graph shown in FIG. 42A to 42D are the same as FIG.
It is a figure which shows the result of the AE process performed based on the graph shown in FIG. FIG. 42A shows an original image (halftone image) before AE processing. FIG. 42B shows histogram data of the average value information LA of the lightness component L * of the original image. The AE processing corrects that the image density of the original image is biased toward the low density side as a whole. FIG. 42C shows an image after AE processing based on the graph shown in FIG. FIG. 42
(D) shows the histogram data after AE processing. Comparing FIGS. 42 (a) and 42 (c), it is understood that the bias of the data distribution is corrected.

<3-4-2-2> Full Color Image AE Process FIG. 43 shows a flowchart of the full color image AE process (step S7013 shown in FIG. 30). here,
The value of the average value information LA of each block is converted into an appropriate value based on the histogram of the average value information LA of the lightness component and the chromaticity component of all the pixel blocks of the original to optimize the density distribution of the image. First, the lightness component L from the hard disk 614
Read out the histogram data of the average value information LA of *,
From the read out histogram data, the minimum value Min, the maximum value Max and the average value Ave of the full-color image are obtained (step S7210). Similarly, the hard disk 61
4, the histogram data of the average value information LA of the lightness components a * and b * is read, and the minimum value Min, the maximum value Max and the average value Ave of the full-color image are obtained from the read histogram data (steps S7211, S72).
12). The AE processing based on the graph in FIG. 40 or FIG. 41 is executed to rewrite the value of the average value information LA of the lightness component L * (step S7213). Similarly, the value of the average value information LA of the chromaticity components a * and b * is rewritten as necessary (steps S7214, S7215).

<3-4-3> Image Editing / Processing Process FIGS. 44 to 48 are flowcharts of the editing / processing process for the attributes of each image. Here, average value information LA for each block, gradation range index LD, and code data φij
The edit / modification process is executed by converting the value of to the planned value. The contents of the editing / processing process are set in the key input process (step S2000). When the value of the mode flag MF for the edit / modification processing set for each image attribute is "1", the following processing is executed. (A) Mono-color conversion process This process targets a full-color image and converts the image into a monochrome image. To convert a full-color image into a monochrome image, set the chromaticity components a * and b * to 0.
However, if the data encoded by the GBTC method is used, the average value information LA of the chromaticity components a * and b * and the value of the gradation width index LD are both converted to 0. Can be realized. This makes the other 4x
It is determined that the image is a full-color image in the attribute determination process without losing the color information of the 4-pixel block.
Only the × 4 pixel block can be converted into a monochrome image. In addition, since encoded data is used, the amount of memory required for conversion can be significantly reduced as compared with the case where 256 gradation data is used for execution. (B) Color conversion process This process is targeted for a monochrome binary image such as a character image and a full-color image. The color conversion process on a black-and-white binary image refers to converting a black character portion and a white background into predetermined lightness and chromaticity. To convert the black and white parts of a black-and-white binary image to the desired lightness and chromaticity,
It is realized by converting the average value information LA and the gradation width information LD of the lightness component L * and the chromaticity components a * and b *. As a result, only the character image portion can be color-converted without changing the color information of other 4 × 4 pixel blocks. Even in the case of a full-color image, color conversion is realized by similar processing. (C) Erase processing This processing is for all images, and for each 4 × 4 pixel block of the selected attribute, its lightness component L
The values of the average value information LA of the chromaticity components a * and b * and the gradation width information LD are both rewritten to 0 and converted into white data. In addition, depending on the setting, each of the attributes other than the selected attribute is 4
It is also possible to perform the trimming operation by converting the data of the × 4 pixel block into white data. (D) dpi conversion processing This processing is intended for a black-and-white multi-valued image and a full-color image, and the code data φij for 4 × 4 pixels is 2 ×
By thinning and storing the data of 2 pixels or 1 × 1 pixel, the image resolution is lowered and the amount of data to be handled is reduced. (E) Negative / Positive Inversion Processing This processing is intended for a black-and-white multi-valued image and a full-color image, and the brightness component L of the corresponding 4 × 4 pixel block.
The value of * is converted into a value obtained by subtracting 256, and the signs of the values of the chromaticity components a * and b * are inverted. This allows
Negative / positive inversion processing of a multi-value gradation image is realized. (F) Background cut processing This processing is intended for a black and white solid image and a color solid image, and the lightness component L of the corresponding 4 × 4 pixel block is used.
The background level is converted by rewriting the average value information LA of *. At this time, the feature amount extraction processing (step S609
0, and the histogram data for the lightness component L * obtained by S6100) is used.

<3-4-3-1> Monochrome Solid Image Editing / Processing Process FIG. 44 is a flowchart showing the monochrome solid image editing / processing process (step S7004 shown in FIG. 29). The value of the mode flag MF13 is "1" for black and white solid images.
Is set to (YES in step S7300).
S) and erase processing (step S7301) are executed. If the value of the mode flag MF14 is set to "1" (YES in step S7302), the negative / positive reversal process (step S7303) is executed. If the value of the mode flag MF15 is set to "1" (YES in step S304), the background cutting process (step S7305) is executed.

<3-4-3-2> Color Solid Image Editing / Processing Process FIG. 45 is a flowchart showing the color solid image editing / processing process (step S7006 shown in FIG. 29). In color solid images, the value of the mode flag MF23 is "
If it is set to "1" (YES in step S7310), the erase process (step S7311) is executed. If the value of the mode flag MF24 is set to "1" (YES in step S7312). , Negative / positive reversal processing (step S7313) If the value of the mode flag MF25 is set to “1” (YES in step S7314), the base cut processing (step S7314)
7315) is executed.

<3-4-3-3> Monochrome Binary Image Editing / Processing Process FIG. 46 is a flowchart showing the monochrome binary image editing / processing process (step S7008 shown in FIG. 29). The value of the mode flag MF32 is "1" in a black and white binary image.
Is set to (YES in step S7320).
S), color conversion processing (step S7321) is executed.
If the value of the mode flag MF33 is set to "1" (YES in step S7322), the erase process (step S7323) is executed. When the value of the mode flag MF34 is set to "1" (step S732)
4, YES), negative / positive reversal processing (step S732)
Execute 5).

<3-4-3-4> Monochrome Multi-valued Image Editing / Processing Processing FIG. 47 is a flowchart showing the black-and-white multi-valued image editing / processing process (step S7011 shown in FIG. 30). The value of the mode flag MF43 is "1" for black and white multi-valued images.
If set to (YES in step S7330).
S) and erase processing (step S7331) are executed. If the value of the mode flag MF44 is set to "1" (YES in step S7332), negative / positive reversal processing (step S7333) is executed. Mode flag
If the value of MF46 is set to "1" (YES in step S7334), dpi conversion processing (step S7)
335) is executed.

<3-4-3-5> Full Color Image Editing / Processing Process FIG. 48 is a flowchart showing the full color image editing / processing process (step S7014 shown in FIG. 30). For full-color images, if the value of the mode flag MF51 is set to "1" (step S73).
YES in 40), mono-color conversion processing (step S73)
41) is executed. When the value of the mode flag MF52 is set to "1" (YES in step S7342)
S) and color conversion processing (step S7343) are executed.
If the value of the mode flag MF53 is set to "1" (YES in step S7344), the erase process (step S7345) is executed. When the value of the mode flag MF54 is set to "1" (step S734
6)), negative / positive reversal processing (step S734)
Execute 7). If the value of the mode flag MF56 is set to "1" (YES in step S7348),
The dpi conversion process (step S7349) is executed.

In the above editing / processing, the values of the average value information LA and the gradation range index LD are changed to the planned values according to the set editing / processing contents based on the attribute discrimination result. The invention is not limited to this, and the value of the decoded data may be changed to a predetermined value.

[0056]

According to the first image processing apparatus of the present invention, the attribute discriminating means can discriminate the attribute of the image of the block based on the average value information and the gradation width index of the encoded data. . Therefore, as compared with the case of discriminating the image attribute from the image data before encoding, the discriminating process can be executed with a simpler operation and a smaller memory. Further, in the second image processing device, the encoded data is further compression-encoded based on the determination result,
The compression ratio of image data can be improved, and the memory required to store code data can be reduced.
Further, in any of the above image processing devices, based on the histogram of the average value information of all blocks, the value of the lightness component of each block or the value of the average value information of the lightness component and the chromaticity component is set to an appropriate value according to the determination result. By converting to, it is possible to optimize the density distribution of the reproduced image without using the data before encoding. As a result, it is possible to save the memory required for the processing and shorten the processing time.
Also, based on the discrimination result, average value information, gradation range index,
By converting the value of the code data φij into a predetermined value, it is possible to edit / process the image without using the data before encoding. As a result, it is possible to save the memory required for the processing and shorten the processing time.

[Brief description of drawings]

FIG. 1 is a diagram for explaining the flow of a general GBTC scheme encoding process.

FIG. 2 is a diagram showing GBTC encoding processing.

FIG. 3 is a sectional view showing the configuration of a digital color copying machine according to this embodiment.

FIG. 4 is a block diagram showing each signal processing executed by the read signal processing unit 106.

FIG. 5 is a diagram showing an L * a * b * color system solid.

FIG. 6 (a) shows that the distribution of the lightness component L * is 0 for each original.
To 255, and (b) and (c) show chromaticity a * and b.
It is a figure which shows the graph used when changing the distribution of each component of * to -127-128 for every manuscript.

7 (a) to (c) are decoded L2 *, a.
The distribution of each data of 2 * and b2 * is the original L * max to L *.
min, a * max to a * min, b * max to b * m
It is a figure which shows the graph used when returning to in.

FIG. 8 is a front view of an operation panel 300.

FIG. 9 is a diagram showing a main routine of a copying process executed by a CPU 611 of the copying machine.

FIG. 10 is a diagram showing a processing flowchart of key input processing (step S2000).

FIG. 11 is a diagram showing a flowchart relating to recompression processing (step S6000) of data encoded by the GBTC method.

FIG. 12 is a diagram showing a flowchart of attribute determination processing (step S6002).

FIG. 13 is a solid image determination process (step S6020).
It is a figure which shows the flowchart of.

FIG. 14 is a color / monochrome discrimination process (step S6).
022, S6026).

FIG. 15 is a binary / multivalue discrimination process (step S602).
It is a figure which shows the flowchart of 8).

FIG. 16 is a monochrome solid image feature amount extraction process (step S
It is a figure which shows the flowchart of 6004).

FIG. 17 is a diagram showing a flowchart of a color solid image feature amount extraction process (step S6006).

FIG. 18: Monochrome binary image feature amount extraction processing (step S
It is a figure which shows the flowchart of 6008).

FIG. 19: Monochrome multi-valued image feature amount extraction processing (step S
It is a figure which shows the flowchart of 6010).

FIG. 20 is a diagram showing a flowchart of full-color image feature amount extraction processing (step S6012).

FIG. 21 is a monochrome solid image recompression process (step S60).
It is a figure which shows the processing flowchart of 05).

FIG. 22 is a conceptual diagram showing a relationship between encoded data and recompressed data formed by a black and white solid image recompression process.

FIG. 23 is a color solid image recompression process (step S6).
It is a figure which shows the processing flowchart of (007).

FIG. 24 is a conceptual diagram showing the relationship between code data and recompressed data formed by color solid image recompression processing.

FIG. 25 is a monochrome binary image recompression process (step S60).
It is a figure which shows the flowchart of 09).

FIG. 26 is a conceptual diagram showing a relationship between code data and recompressed data formed by the monochrome binary image recompression process.

FIG. 27 is a recompression process for a monochrome multi-valued image (step S6).
It is a figure which shows the flowchart of 011).

FIG. 28 is a conceptual diagram showing the relationship between coded data and recompressed data formed by recompressing a monochrome multi-valued image.

FIG. 29 is a diagram showing a flowchart of recompression / expansion processing (step S7000).

FIG. 30 is a diagram showing a flowchart of recompression / expansion processing (step S7000).

FIG. 31 is a diagram showing a flowchart of decompression processing (step S7003) from recompression of a monochrome solid image.

FIG. 32 is a diagram showing recompressed data in the upper part and code data obtained by expanding the recompressed data in the lower part.

FIG. 33 is a diagram showing a flowchart of decompression processing from color solid image recompression (step S7005).

FIG. 34 is a diagram showing recompressed data in the upper part and code data obtained by expanding the recompressed data in the lower part.

FIG. 35 is a diagram showing a flowchart of decompression processing (step S7007) from monochrome binary image recompressed data.

FIG. 36 is a diagram showing recompressed data in the upper part and code data obtained by expanding the recompressed data in the lower part.

[Fig. 37] Fig. 37 is a diagram illustrating a flowchart of a decompression process (step S7009) from recompression of a monochrome multi-valued image.

FIG. 38 is a diagram showing recompressed data in the upper part, and code data obtained by expanding the recompressed data in the lower part.

FIG. 39 is a diagram showing a processing flowchart of AE processing (step S7010) for a monochrome multi-valued image.

FIG. 40 is a graph showing a relationship between average value information LA and average value information LA ′ after AE processing.

FIG. 41 is a graph showing a relationship between average value information LA and another average value information LA ″ after AE processing.

42A shows a document image (halftone image) before AE processing, FIG. 42B shows histogram data of average value information LA of the lightness component L * of the image shown in FIG.
(C) shows the image after AE processing, (d) shows the histogram data of the original image after AE processing.

FIG. 43 is a diagram showing a flowchart of full-color image AE / color adjustment processing (step S7013).

FIG. 44: Monochrome solid image editing / processing (step S
It is a figure which shows the flowchart of (7004).

FIG. 45 is a diagram showing a flowchart of a color solid image editing / modifying process (step S7006).

FIG. 46: Black and white binary image editing / processing (step S
It is a figure which shows the flowchart of (7008).

FIG. 47: Black-and-white multi-valued image editing / modification processing (step S
It is a figure which shows the flowchart of 7011).

FIG. 48 is a diagram showing a flowchart of full-color image editing / modification processing (step S7014).

[Explanation of symbols]

 Reference numeral 106 ... Read signal processing unit 300 ... Operation panel 601 ... Reading device color correction processing unit 602 ... Color space conversion processing unit 603 ... Color space optimization processing unit 604 ... Encoding / decoding processing unit 605 ... Color space inverse optimization processing Part 606 ... Color space inverse conversion processing part 607 ... Reflection / density conversion processing part 608 ... Masking processing part 610 ... Compressed image memory 611 ... CPU 614 ... Hard disk

Claims (5)

[Claims] 1. A data conversion processing unit for converting RGB image data of an original into lightness component data and chromaticity component data, and the lightness component and chromaticity component data is divided into blocks each having a predetermined pixel matrix. However, for each block, the parameter P1 determined from the data in the block
Average value information obtained by dividing the sum of the average value Q1 of the data of the following values and the average value Q4 of the data of the values of the parameter P2 (where P1 <P2) or more, and the average value Based on Q4 and the gradation width index that is the difference between the average values Q1, the data of each pixel in the block is quantized within the range of the gradation distribution in the block with a gradation level smaller than that of the data. An encoding processing unit that encodes encoded data, and an image to which the block belongs, based on the average value information of the lightness component and chromaticity component, the gradation width index, and the value of the encoded data for the encoded block An attribute discrimination processing unit that discriminates the attribute, a storage unit that stores, for each block, average value information, a gradation range index, code data, and an attribute determination result, and average value information that is stored in the storage unit. And the gradation Based on the index, an image processing apparatus characterized by comprising a decoding unit for decoding the encoded data in units of blocks. 2. A data conversion processing unit for converting RGB image data of an original into lightness component data and chromaticity component data, and the lightness component and chromaticity component data are divided into blocks each having a predetermined pixel matrix. Then, for each block, a parameter P determined by the data in the block
Average value information obtained by dividing the sum of the average value Q1 of data having a value of 1 or less and the average value Q4 of data having a value of the parameter P2 (provided that P1 <P2) or more into two equal parts, and the above average Based on the gradation width index which is the difference between the value Q4 and the average value Q1, the data of each pixel in the block is quantized within the range of the gradation distribution in the block with a gradation level smaller than that of the data. An encoding processing unit that encodes encoded data, and an image to which the block belongs, based on the average value information of the lightness component and chromaticity component, the gradation width index, and the value of the encoded data for the encoded block An attribute discrimination processing unit for discriminating attributes, average value information of lightness component and chromaticity component, gradation range index,
A compression processing unit that executes a compression process that reduces data that can always be identified based on the attribute of the determined image among the code data, a data unit after the compression process, and a storage unit that stores the attribute determination result, The compressed data decompression processing unit that reads the data after the compression processing and the attribute determination result stored in the storage unit and restores the data reduced in the compression processing based on the attribute determination result, and the data decompression processing unit that decompresses the data. An image processing apparatus comprising: a decoding processing unit that decodes coded data in blocks on the basis of average value information of lightness components and chromaticity components and a gradation width index. 3. The image processing apparatus according to claim 1 or 2, wherein the attribute discrimination processing unit determines that the gradation width indexes of the brightness component and the chromaticity component of the block to be discriminated are all equal to or less than a predetermined value. In some cases, if the determination process determines that the block belongs to a solid image, and if even one value of the tone width index of the lightness component and the chromaticity component has a value equal to or greater than the predetermined value, the block image is If the determination process of determining that the block belongs to an image having two or more gradations and the average value information of the chromaticity components of the blocks to be determined are both below a predetermined value different from the above, the block is black and white. Determination processing for determining that the image belongs to the image,
When the value of the average value information of the chromaticity component is equal to or more than a predetermined value different from the above, the block is determined to belong to the color image, and the code assigned to the brightness component of the block to be determined. When the data value is polarized, it is determined that the block belongs to a binary image, and when the data is not polarized, at least the determination process of determining that the block belongs to a multi-valued image is performed. An image processing apparatus, which executes one determination process. 4. The image processing apparatus according to claim 1, further comprising the step of: before decoding the coded data in block units in the coding processing unit. Depending on the attribute, the histogram of the average value information of the brightness component of all the blocks of the attribute, or the brightness component and the chromaticity component is obtained, and the value of the average value information of each block is converted to an appropriate value based on the obtained histogram. An image processing apparatus comprising: an average value information conversion processing unit that 5. The image processing apparatus according to any one of claims 1 to 3, further comprising: average value information of a lightness component and a chromaticity component, a gradation width index, in a block unit from a storage unit. Read the coded data and the attribute discrimination result, and based on the read attribute discrimination result,
An image processing apparatus comprising: an edit / processing unit that changes the average value information of the corresponding component, the gradation width index, and the value of the code data to a predetermined value.