JP2023037968A - Image processing apparatus, control method of the same, and program - Google Patents

Abstract

To provide an image processing apparatus capable of reducing a calculation load for performing image processing of separating an image signal into a character region and a non-character region.SOLUTION: In an MFP 100, a CPU 201 executes a predetermined program, thereby extracting a rectangular image made of a target pixel and pixels therearound from a document image scanned by a scanner 120, obtains a standard deviation of a signal value of the extracted rectangular image, and generates image region information indicating whether the target pixel is a character or not, for example, on the basis of the standard deviation.SELECTED DRAWING: Figure 5

Description

The present invention relates to an image processing apparatus and its control method and program.

When copying a document containing mixed characters and halftone images with a digital color multifunction printer (MFP), various processing is applied to the image signal input from the scanner unit in order to obtain a higher quality print. There is something. For example, in order to output a character portion sharply, filter processing for emphasizing the contour is performed on the image signal. On the other hand, if a halftone dot image portion is subjected to a filtering process for emphasizing the contour, moiré will occur in the processed image. Therefore, in many MFPs, all image signals are separated pixel by pixel into character areas and non-character areas including halftone dot areas, and appropriate filter processing is performed for each separated area.

An example of a method of separating an image signal into a character area and a non-character area on a pixel-by-pixel basis is disclosed in Japanese Unexamined Patent Application Publication No. 2002-200512. In Patent Document 1, first, an original image is binarized, and then all pixels are divided into character pixels and non-character pixels by edge determination. A pattern matching process is then used to detect the similarity of the character pixels to the halftone screen used in the printed matter. As a result, since there is a high possibility that text pixels with high similarity are pixels forming halftone dots, the classification is changed from text pixels to non-text pixels.

JP-A-2006-5657

When pattern matching processing is used as in the technique disclosed in Patent Document 1, it is necessary to determine various patterns at the same time. there is However, the dedicated ASIC has the problem of lack of functional expandability and high cost.

As a solution to this problem, a configuration is conceivable in which a CPU or GPU, which performs various controls in the image forming apparatus, executes processing for separating an image signal into character areas and non-character areas without using a dedicated ASIC. However, the CPU and GPU need to control the hardware of the MFP and perform rendering processing and the like. Therefore, if an image signal is separated into a character area and a non-character area by pixels while executing these controls and processes to obtain a high-quality printed matter, the output speed of the printed matter drops significantly. Therefore, there is a demand for a technique for reducing the computational load on the CPU and GPU in image processing for separating an image signal into character areas and non-character areas.

SUMMARY OF THE INVENTION It is an object of the present invention to provide an image processing apparatus capable of reducing the computation load when a CPU or GPU performs image processing for separating an image signal into character areas and non-character areas.

An image processing apparatus according to the present invention is an image processing apparatus for processing a document image read by a document reading apparatus, and is a rectangular image formed of pixels selected from the document image and pixels surrounding the selected pixels. and generating means for generating image area information indicated by the selected pixel based on the statistical value.

According to the present invention, it is possible to reduce the calculation load in image processing in which a CPU or GPU separates an image signal into character areas and non-character areas.

1 is a diagram showing a schematic configuration of an MFP (multifunction peripheral) according to an embodiment; FIG. 3 is a block diagram of a controller mounted on the MFP; FIG. It is the figure which illustrated the program stored in HDD etc. of a controller. 4 is a flowchart of copy processing executed by the MFP; 5 is a flowchart of image area information generation processing in the first embodiment; It is a schematic diagram explaining the determination method of a fine line. FIG. 10 is a diagram showing an example of the configuration and frequency characteristics of a filter used in S504; FIG. 4 is a diagram schematically showing signals before and after smoothing processing of each pixel in a rectangular image centered on a pixel of interest; 4 is a flowchart of descreen processing in the first embodiment; FIG. 10 is a diagram showing an example of the configuration and frequency characteristics of a filter used in S902; 6 is a flowchart of edge enhancement processing in the first embodiment; FIG. 10 is a diagram showing an example of the configuration and frequency characteristics of a filter used in S1103; 6 is a flowchart of print data synthesis processing in the first embodiment; 10 is a flowchart of image area information generation processing in the second embodiment; 9 is a flowchart of descreen processing in the second embodiment; 9 is a flowchart of edge enhancement processing in the second embodiment; 9 is a flowchart of print data synthesizing processing in the second embodiment;

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Here, a configuration in which the image processing apparatus according to the present invention is embodied as an MFP (complex machine) will be described.

<First embodiment>
FIG. 1 is a diagram showing a schematic configuration of an MFP 100 (complex machine) according to the embodiment. The MFP 100 includes a document platen 110, a scanner 120, a printer 130, and an operation unit 122, and can execute a copy function, a scan function, and a color printing function (color image forming function). Note that the mechanical configuration of the MFP 100 may be a well-known configuration, and detailed description thereof will be omitted.

MFP 100 implements controller 140 . FIG. 2 is a block diagram of the controller 140. As shown in FIG. The controller 140 has a CPU 201 , a ROM 202 , a RAM 203 , an HDD 204 and a GPU 205 mounted so as to be able to communicate bidirectionally via an internal bus 210 . The CPU 201 loads predetermined programs stored in the ROM 202 and the HDD 204 into the RAM 203, thereby controlling the operation of each unit of the MFP 100 and executing predetermined processing.

The controller 140 is connected to the printer 130, the scanner 120, the network I/F 207, and the operation unit 122 via the internal bus 210 and circuit patterns and cables (not shown). With such a configuration, the scanner 120 reads an original image, the CPU 201 executes a predetermined program on the read image, performs predetermined image processing, and the printer 130 prints it on a recording material such as paper, thereby providing a copy function. can be realized. At that time, by having the GPU 205 perform the image processing instead of the CPU 201, the image processing can be speeded up. In the following description, the expression that the CPU 201 executes image processing includes that the GPU 205 executes image processing. The printing method of the printer 130 is not limited, and for example, a known method such as an electrophotographic method or an inkjet method can be used. there is

FIG. 3 is a diagram exemplifying programs stored in the ROM 202 and the HDD 204. As shown in FIG. Various programs shown in FIG. 3 are executed by the CPU 201 and the GPU 205 . Various programs shown in FIG. 3 will be described as appropriate next together with the description of the flowchart of FIG.

FIG. 4 is a flowchart of copy processing executed by MFP 100 . Each process (step) denoted by S number in FIG. 4 is realized by the CPU 201 loading a predetermined program stored in the ROM 202 and/or the HDD 204 into the RAM 203 and controlling the operation of each part of the MFP 100 .

The CPU 201 waits while executing the operation unit acceptance program 309 (see FIG. 3). For example, when a copy button (not shown) of the operation unit 122 is pressed, the CPU 201 receives a copy processing execution command from the operation unit 122 and executes the copy processing program 310 (see FIG. 3).

In S401, the CPU 201 executes the document reading program 302 (see FIG. 3). As a result, the scanner 120 reads the document 112 placed on the document platen 110 , generates RGB color image data, and temporarily stores the generated image data in the RAM 203 .

Note that the scanner 120 sequentially reads the document 112 from the upper end to the lower end. Although the number of originals to be read by the scanner 120 is not limited, the flowchart in FIG. 4 shows the processing for one original 112 for convenience of explanation. When a plurality of documents 112 are read in S401, the processing of S402 to S408, which will be described later, is performed for each document.

In S402, the CPU 201 selects (extracts) a pixel of interest from the image data stored in the RAM 203 in S401. It should be noted that the CPU 201 can asynchronously proceed with the subsequent processing for each pixel at the same time. For example, while processing the first pixel (x,y)=(0,0), processing of the second pixel (x,y)=(0,1) can begin. However, if the calculation results of the pixels surrounding the pixel of interest are required during the processing, the processing of the pixel of interest is suspended until the calculation of the surrounding pixels is completed.

In S<b>403 , the CPU 201 executes the image area information generation processing program 303 (see FIG. 3 ) to generate image area information of the pixel of interest selected in S<b>402 and temporarily stores the generated image area information in the RAM 203 . Details of the processing of S403 will be described later. Note that, in the first embodiment, it is mainly assumed that the copy processing of the same size is performed, and therefore, it is necessary to perform all the processing for each pixel.

In S404, the CPU 201 executes the descreen processing program 304 (see FIG. 3) to perform descreen processing on the pixel of interest selected in S402 using the image area information generated in S403. Details of the processing of S404 will be described later. In S405, the CPU 201 executes the edge enhancement processing program 305 (see FIG. 3), executes edge enhancement processing using the image generated in S404 and the image area information generated in S403, and stores the obtained image data in the RAM 203. Save to Details of the processing of S405 will be described later.

In step S406, the CPU 201 executes the print data composition processing program 308 (see FIG. 3), and executes print data composition processing using the image data and image area information generated in the previous processing. Details of the processing of S406 will be described later. In S407, the CPU 201 determines whether the processing of S402 to S406 has been completed for all pixels of one page. When the CPU 201 determines that the processing for one page has not been completed (No in S407), the process returns to S402, and when it determines that the processing for one page has been completed (Yes in S407), the process proceeds to S408. proceed to In S408, the CPU 201 executes the print processing program 311 (see FIG. 3), transfers the print data generated in S407 to the printer 130, and the printer 130 performs print processing on paper, thereby ending this processing.

Next, image area information generation processing in S403 will be described. FIG. 5 is a flowchart of image area information generation processing in S403 in the first embodiment. In S<b>501 , the CPU 201 selects a rectangular area including the target pixel determined in S<b>402 and surrounding pixels as reference pixels, and acquires an RGB color rectangular image from the RAM 203 . The CPU 201 selects, for example, a rectangular area of 5 vertical pixels×5 horizontal pixels centered on the pixel of interest.

In S502, the CPU 201 converts the RGB color rectangular image acquired in S501 into a grayscale rectangular image. As for the method of conversion to gray scale, for example, it is sufficient to extract only the 8-bit signal of the G (green) channel from the RGB 24-bit signal of each pixel. In step S<b>503 , the CPU 201 determines whether the target pixel is a pixel forming a thin line in the grayscale rectangular image. The reason why the process of S503 is performed is that if the smoothing process of S504 is performed on the pixels constituting the thin line, it may become difficult to distinguish between the halftone dot portion and the thin line portion in the standard deviation calculation of S505. This is to avoid smoothing the pixels.

Here, an example of a thin line determination method will be described. FIG. 6 is a schematic diagram for explaining a fine line determination method. FIG. 6A shows a rectangular image of 5 vertical pixels×5 horizontal pixels centered on the pixel of interest. Continuity of density is detected in the directions of arrows α, β, γ, and δ shown in FIG. 6(a). Continuity detection can be performed, for example, by simply binarizing each pixel (0 (white), 255 (black)) and determining whether or not it is all black pixels in each direction. . If all the pixels are black in a plurality of directions of the arrows α, β, γ, and δ, for example, in both directions of the arrows α and β, it is determined that the pixel of interest is not a pixel forming a fine line. .

6B and 6C are diagrams showing examples of rectangular images. In the rectangular image of FIG. 6B, the pixel of interest is black, and the black pixels are continuously arranged in the direction of the arrow β, so the pixel of interest is determined to be part of a fine line. On the other hand, in the rectangular image shown in FIG. 6C, black pixels are continuously arranged in the directions of the arrows α and β. is determined.

If the CPU 201 determines that the pixel of interest does not form a thin line (No in S503), it advances the process to S504. proceed to

In S504, the CPU 201 performs smoothing processing using a low-pass filter on the rectangular image in which it is determined that the pixel of interest does not form a fine line. FIG. 7 is a schematic diagram for explaining the smoothing process in S504. FIG. 7A is a diagram showing an example of a low-pass filter used in S504. FIG. 7(b) is a diagram showing the frequency characteristics of the low-pass filter of FIG. 7(a). In S504, the low-pass filter shown in FIG. 7A attenuates the high frequency component with the frequency characteristics shown in FIG. 7B. However, according to the determination in S503, the smoothing process in S504 is not performed on the pixels forming the thin line. After completing the process of S504, the CPU 201 advances the process to S505.

In S505, the CPU 201 calculates the statistic value of the rectangular image, and sets the calculated statistic value as the statistic value of the pixel of interest at that time. As the statistical value, for example, the standard deviation is effective, so the standard deviation is used as the statistical value in the first embodiment. Note that the maximum and minimum values of standard deviation are determined in advance and stored in the ROM 202 or HDD 204 . The standard deviation of a rectangular image of 5 pixels long by 5 pixels wide is calculated by the following equation 1. Then, the signal is normalized to a signal having a signal value (pixel value) of 0 to 255 according to Equation 2 below.

In Equations 1 and 2 below, gray(i, j) is the signal value of each pixel in the rectangular image, and the overlined gray is the average value of the pixel signal values in the rectangular image. Also, max is the maximum pre-determined standard deviation, min is the minimum pre-determined standard deviation, and flag is the normalized standard deviation.

The maximum value (max) of the standard deviation determined in advance is desirably determined using the maximum value of the standard deviation when the pixels forming the characters are read by the scanner 120 . In addition, the minimum value (min) of the standard deviation determined in advance is determined using the minimum value when the scanner 120 reads the background portion of the document (paper color (generally white)). is desirable.

Here, the processing of S503 to S505 will be described with reference to FIG. FIG. 8(a) is a diagram simply showing an example of a signal in a rectangular image when the pixel of interest is a halftone dot portion of an original. FIG. 8(b) is a diagram simply showing the state after the rectangular image in FIG. 8(a) is smoothed in S504. As shown in FIG. 8B, the smoothing process increases the number of halftones, and therefore the standard deviation calculated in S505 tends to decrease. As a result, it becomes easier to determine that it is a non-character.

FIG. 8(c) is a diagram simply showing an example of a signal in a rectangular image when the pixel of interest is an edge portion of characters or lines of a document. FIG. 8(d) is a diagram simply showing the state after the rectangular image in FIG. 8(c) is smoothed in S504. As shown in FIG. 8D, the smoothing process increases the number of high-density pixels and low-density pixels, so that the standard deviation calculated in S505 tends to increase. As a result, it becomes easy to distinguish that it is a character.

FIG. 8(e) is a diagram simply showing an example of a signal in a rectangular image when the pixel of interest is an intersection of characters or lines of a document. FIG. 8(f) is a diagram simply showing the state after performing the smoothing processing of S504 on the rectangular image of FIG. 8(e). As shown in FIG. 8(f), the smoothing process increases the number of high-density pixels and low-density pixels, so the standard deviation calculated in S505 tends to increase. As a result, it becomes easy to distinguish that it is a character. However, in S503, pixels forming a line may be erroneously determined, and the standard deviation in S505 may be calculated with the signals of the rectangular image shown in FIG. 8E. Even in that case, since there are many high-density pixels and low-density pixels, it is likely to be determined to be a character.

FIG. 8(g) is a diagram simply showing an example of a signal in a rectangular image when the pixel of interest is a fine line of the document. FIG. 8(h) is a diagram showing a state after performing the smoothing processing of S504 on the rectangular image of FIG. 8(g). As shown in FIG. 8(h), the smoothing process increases the number of halftone pixels, so that the standard deviation calculated in S505 tends to decrease. As a result, it becomes easier to determine that it is a non-character. However, since it can be determined in S503 that the rectangular image in FIG. 8G is a pixel forming a fine line, the signal of the rectangular image in FIG. Then, in S506, the standard deviation is calculated. Even in this case, since high-density pixels and low-density pixels are mixed in FIG.

Now, in S506 after the processing of S505, the CPU 201 performs fattening processing. The thickening process is performed by calculating the standard deviations obtained in S505 for each of the nine pixels of the pixel of interest and the pixels adjacent to the pixel of interest, and then calculating the standard deviation of the pixel of interest as the maximum standard deviation among the nine pixels. This is the process of replacing with the value of FIG. 6D is a diagram schematically showing the pixel of interest for which the processing up to S505 has been completed and the nine pixels surrounding it. The value of the standard deviation of the pixel of interest is replaced with the maximum value among the standard deviations of these nine pixels. The CPU 201 saves the standard deviation obtained by the processing up to S506 in the RAM 203 as image area information of the coordinates of the pixel of interest, and terminates this processing.

Next, the descreen process of S404 will be described. FIG. 9 is a flowchart of descreen processing in S404 in the first embodiment. Since the processing of S901 is the same as the processing of S501 (see FIG. 5), description thereof is omitted. In S902, the CPU 201 performs descreen processing using a digital filter on each channel (RGB) of the RGB color rectangular image acquired in S901. FIG. 10A is a diagram showing an example of a digital filter used in S902. The descreening process in S902 is a convolution operation using a digital filter.

FIG. 10(b) is a diagram showing frequency characteristics of the digital filter of FIG. 10(a). Descreen processing is performed by applying a digital filter to each channel of the RGB color rectangular image. If the RGB color rectangular image acquired in S901 is image data forming the halftone dot portion of the document, the high frequency component of the halftone dot is removed by the descreening process.

In S<b>903 , the CPU 201 acquires image area information of the same coordinates as the target pixel from the RAM 203 . In S904, the CPU 201 calculates the signal value of the pixel of interest acquired in S901 and the output signal of the pixel of interest after the descreening process in S902 using the following equations 3, 4, and 5 using the image area information read in S903. . As a result, output signals (R(x, y), G(x, y), B(x, y)) of respective colors of the pixel of interest after combining the image area information are obtained. The CPU 201 saves the obtained output signal in the RAM 203 and terminates this process.

Note that (x, y) are the coordinates of the pixel of interest in Equations 3 to 5 below. Also, R0, G0, and B0 are the signal values of the respective colors of the pixel of interest acquired in S901. R1, G1, and B1 are the signal values of the respective colors of the pixel of interest after the descreening process is performed in S902, and flag is the image area information of the pixel of interest read in S903.

Next, edge enhancement processing in S405 will be described. FIG. 11 is a flowchart of edge enhancement processing in S405 in the first embodiment. In S1101, the CPU 201 selects a rectangular area including the pixel of interest determined in S402 and its surrounding pixels as reference pixels, and acquires a descreen-processed RGB color rectangular image from the RAM 203 in S404. The CPU 201 selects, for example, a rectangular area of 5 vertical pixels×5 horizontal pixels centered on the pixel of interest.

In S1102, the CPU 201 converts the RGB color rectangular image acquired in S1101 into a color space such as CIE-L ^* a ^* b ^* that can be separated into lightness and color difference components. In S1103, the CPU 201 applies a character filter to the lightness (L of CIE-L ^* a ^* b ^* ) generated in S1102, performs a convolution operation, and sets the obtained result as L0. FIG. 12A is a diagram showing an example of a character filter used in S1103. FIG. 12(b) is a diagram showing the frequency characteristics of the character filter of FIG. 12(a). The character filter has a characteristic of emphasizing edge components (increasing the power of the intermediate band), and by convolution operation with an L (lightness) image, an L (lightness) image with a sharp impression is generated.

In S1104, the CPU 201 applies the photographic filter to the brightness generated in S1102, performs convolution operation, and sets the obtained result as L1. FIG. 12C is a diagram showing an example of a photographic filter used in S1104. FIG. 12(d) is a diagram showing the frequency characteristics of the photographic filter of FIG. 12(c). A photographic filter has a characteristic of strongly removing noise components, and by suppressing changes up to a certain band and attenuating high frequency components, an L (lightness) image with a smooth impression is generated.

In S<b>1105 , the CPU 201 acquires from the RAM 203 image area information (flag(x, y)) of the same coordinates as the pixel of interest (x, y). In S1106, the CPU 201 uses the lightness (L0) after text filter processing, the lightness (L1) after photo filter processing, and the image area information (flag (x, y)) acquired in S1103 to S1105. Then, the lightness signal L of the pixel of interest is calculated by the following equation (6).

In S1107, the CPU 201 generates an RGB color image from the CIE-L ^* a ^* b ^* lightness signal L generated in S1106 and the color image of the separated color difference components (a ^* b ^* ), and converts the generated RGB color image. is stored in the RAM 203, and this process is terminated.

Next, the print data combining process in S406 will be described. FIG. 13 is a flowchart of print data synthesis processing in S406 in the first embodiment. In S1301, the CPU 201 reads from the RAM 203 the RGB color image of the pixel of interest generated in S405.

In S1302, the CPU 201 uses the color conversion processing program 306 (see FIG. 3) to convert the RGB color image acquired in S1301 into colors of color materials for printing, such as C (cyan), M (magenta), Y (yellow) and K (black) signals. In general, when performing color conversion from RGB to CMYK, color conversion using a three-dimensional matrix and interpolation calculation is used. At this time, in S1302, CMYK color signal values are generated assuming that the pixel of interest is a character portion of the document. For example, if the signal value is (R, G, B)=(0, 0, 0), then (C, M, Y, K)=(0, 0, 0, 255). By adjusting the character pixels so that they are composed only of black toner, the amount of toner used for expressing black can be reduced.

In S1303, the CPU 201 performs the same processing as in S1302, but here, CMYK color signal values are generated assuming that the pixel of interest is a non-character portion of the document. For example, if the signal value is (R, G, B)=(0, 0, 0), then (C, M, Y, K)=(30, 30, 30, 200). Even in black pixels, by mixing color toners other than black toner, it is possible to express deeper and richer black as used in photo printing.

In S<b>1304 , the CPU 201 acquires image area information of the same coordinates as the target pixel from the RAM 203 . In S1305, the CPU 201 synthesizes the character CMYK image generated in S1302 and the non-character CMYK image generated in S1303 using the following equations 7 to 10 to obtain color material signals. Note that (x, y) are the coordinates of the pixel of interest, C, M, Y, and K are respectively synthesized colorant signals, C0, M0, Y0, and K0 are respectively colorant signals generated in S1302, C1, M1, Y1, and K1 are color material signals generated in S1303. Also, flag is the image area information acquired in S1304.

In step S<b>1306 , the CPU 201 uses the halftone processing program 307 (see FIG. 3 ) to convert the CMYK image generated in step S<b>1305 to a pseudo-gradation halftone to a bit depth (for example, 1 bit) printable by the printer 130 . Convert to image. A conversion method to a halftone image is not limited, and known methods such as an error diffusion method and a dither method can be used. The CPU 201 terminates this process after the process of S1306.

As described above, according to the first embodiment, it is possible to reduce the calculation load in image processing for separating an image signal into character areas and non-character areas. This allows the CPU 201 and GPU 205 to execute high-speed, high-quality print processing while executing overall control in the MFP 100 such as hardware control and rendering processing.

<Second embodiment>
If the image area information for each pixel is stored in the RAM 203, it cannot be used for highly efficient lossy compression such as JPEG. This is because there is a possibility that some of the character pixels of the original will be processed as non-characters due to noise generated by irreversible compression, resulting in a significant deterioration in image quality.

Therefore, in the second embodiment, a configuration for acquiring and handling image area information as binary information will be described. Note that in the second embodiment, processing for handling image area information as binary information will be mainly described, and description of processing common to the first embodiment will be omitted. Overall processing, taking copy processing as an example, conforms to the flowchart in FIG. In the second embodiment, the processes of S403 to S406 among the processes in the flowchart of FIG. 4 are different from those in the first embodiment, so these processes will be described below.

FIG. 14 is a flowchart of S403 (image area information generation processing) in the second embodiment. Since each process of S1401 to S1406 is the same as the process of S501 to S506 in the flow chart of FIG. 5, description thereof will be omitted. In S1407, the CPU 201 determines whether the value of the standard deviation of the pixel of interest after the calculation in S1406 is greater than a predetermined threshold. If the CPU 201 determines that the standard deviation value of the pixel of interest is greater than the predetermined threshold value (Yes in S1407), the CPU 201 advances the process to S1408. (No in S1407), the process proceeds to S1409.

In S1408, the CPU 201 stores a flag indicating text information in the image area information, and terminates this processing. In S1409, the CPU 201 stores a flag indicating non-character information in the image area information, and terminates this processing. By the above processing, binary values of text/non-text for one page are generated in the image area information generation processing of S403.

FIG. 15 is a flowchart of S404 (descreen processing) in the second embodiment. In S<b>1501 , the CPU 201 acquires image area information of the pixel of interest from the RAM 203 . In S1502, the CPU 201 determines whether the image area information acquired in S1501 indicates non-character. The processing in S1502 is performed based on the flags attached to the image area information in S1408 and S1409. If the CPU 201 determines that the image area information indicates non-characters (Yes in S1502), the process advances to S1503 to determine that the image area information does not indicate non-characters (represents characters). If so (No in S1502), this process is terminated. The processing of S1503 and S1504 is the same as the processing of S901 and S902 in the flowchart of FIG. 9, respectively, so the description is omitted.

FIG. 16 is a flowchart of S405 (edge enhancement processing) in the second embodiment. Since the processing of S1601 to S1603 is the same as the processing of S1101 to S1103 in the flowchart of FIG. 11, the description is omitted. In S1604, the CPU 201 determines whether the image area information of the target pixel acquired in S1603 indicates non-character. If the CPU 201 determines that the image area information indicates non-characters (Yes in S1604), the process advances to S1605 to determine that the image area information does not indicate non-characters (represents characters). If so (No in S1604), the process advances to S1606. Since the processing of S1605, S1606 and S1607 is the same as the processing of S1104, S1103 and S1107 in the flowchart of FIG. 11, the description thereof is omitted.

FIG. 17 is a flowchart of print data synthesizing processing in the second embodiment. Since the processing of S1701 and S1702 is the same as the processing of S1301 and S1304 in the flowchart of FIG. 13, the description is omitted. In S1703, the CPU 201 determines whether the image area information of the target pixel acquired in S1702 indicates non-character. If the CPU 201 determines that the image area information indicates non-characters (Yes in S1703), the process advances to S1704 to determine that the image area information does not indicate non-characters (represents characters). If so (No in S1703), the process advances to S1705. The processing of S1704, S1705, and S1706 is the same as the processing of S1303, S1302, and S1306 in the flowchart of FIG. 13, so description thereof will be omitted. Even in such a second embodiment, the same effect as in the first embodiment can be obtained.

Although the present invention has been described in detail based on its preferred embodiments, the present invention is not limited to these specific embodiments, and various forms without departing from the gist of the present invention can be applied to the present invention. included. Furthermore, each embodiment described above merely shows one embodiment of the present invention, and it is also possible to combine each embodiment as appropriate.

The present invention supplies a program that implements one or more functions of the above-described embodiments to a system or apparatus via a network or a storage medium, and one or more processors in the computer of the system or apparatus reads and executes the program. It can also be realized by processing to

120 Scanner 201 CPU
202 ROMs
204 HDDs
205 GPUs
303 image area information generation processing program 304 descreen processing program 305 edge enhancement processing program 306 color conversion processing program 308 print data synthesis processing program

Claims (15)

An image processing device for processing a document image read by a document reading device,
a first computing means for computing statistical values of a rectangular image composed of pixels selected from the document image and pixels surrounding the selected pixels;
and generating means for generating image area information indicated by the selected pixels based on the statistical value. 2. The image processing apparatus according to claim 1, wherein said statistical value is standard deviation. 3. The image processing according to claim 2, wherein the standard deviation is obtained using a signal value of each pixel forming the rectangular image and an average value of signal values of each pixel forming the rectangular image. Device. a first conversion means for converting the rectangular image to grayscale;
determining means for determining whether the selected pixel is a pixel forming a thin line in the rectangular image;
smoothing processing means for performing smoothing processing on the rectangular image converted to grayscale including the selected pixels when the selected pixels do not form a thin line;
When the selected pixels constitute a thin line, the first computing means computes the standard deviation using a signal value not subjected to smoothing processing in a rectangular image containing the selected pixels, and 3. When the selected pixels do not constitute a thin line, the standard deviation is calculated using the signal value after the smoothing process in the rectangular image including the selected pixels. The image processing device according to . The generating means is
replacement means for replacing the statistical value of the selected pixel with the maximum value of the statistical values of the selected pixel and pixels adjacent to the selected pixel;
5. The image processing apparatus according to claim 1, further comprising storage means for storing the statistical value replaced by said replacement means as image area information of said selected pixel. descreen processing means for performing descreen processing on the rectangular image using a digital filter;
a second computing means for computing the signal value of each color of the selected pixel using the original signal value, the descreened signal value, and the image area information of the selected pixel; 6. The image processing apparatus according to claim 5, further comprising: edge enhancement means for performing edge enhancement processing on the descreened rectangular image;
The edge enhancement means is
Separating means for separating the rectangular image into lightness and color difference component signals;
a third computing means for computing the brightness signal of the selected pixel by applying a character filter to the brightness signal generated by the separating means;
fourth computing means for computing the brightness signal of the selected pixel by applying a photographic filter to the brightness signal generated by the separation means;
7. The image forming apparatus according to claim 6, further comprising: fifth computing means for computing the brightness signal of the selected pixel from the brightness signals obtained by the third computing means and the fourth computing means. Image processing device. conversion means for converting the lightness signal obtained by the fifth calculation means into a color material signal;
8. The image processing apparatus according to claim 7, further comprising synthesizing means for synthesizing print data from said image area information and said colorant signal. The storage means stores the image area information of the selected pixel as character information when the statistical value replaced by the replacing means is larger than a predetermined threshold value, and the statistical value replaced by the replacing means is a predetermined value. 6. The image processing apparatus according to claim 5, wherein the non-character information is stored when the value is equal to or less than the threshold value of . descreen processing means for performing descreen processing on the rectangular image including the selected pixels using a digital filter when the image area information of the selected pixels is the non-character information. 10. The image processing apparatus according to claim 9. edge enhancement means for performing edge enhancement processing on the descreened rectangular image;
The edge enhancement means is
Separating means for separating the rectangular image into lightness and color difference component signals;
When the image area information of the selected pixel is the character information, the brightness of the selected pixel is obtained by applying a character filter to the brightness signal of the rectangular image including the selected pixel. a third computing means for computing a signal;
When the image area information of the selected pixels is the non-character information, applying a photographic filter to a lightness signal of a rectangular image including the selected pixels a fourth computing means for computing the brightness signal;
5. A fifth computing means for computing the brightness signal of the selected pixel from the brightness signal obtained by the third computing means and the fourth computing means and the statistical value. 11. The image processing apparatus according to 10. conversion means for converting the lightness signal obtained by the fifth calculation means into a color material signal;
12. The image processing apparatus according to claim 11, further comprising synthesizing means for synthesizing print data from said image area information and said colorant signal. 13. The image processing apparatus according to any one of claims 1 to 12, wherein said first computing means and said generating means are a CPU or a GPU that controls the entire image processing apparatus. A control method for an image processing device,
calculating statistical values of a rectangular image composed of pixels selected from a document image read by a document reading device and pixels surrounding the selected pixels;
and generating image area information indicated by the selected pixels based on the statistical value. A program that causes a computer to function as each means of the image processing apparatus according to any one of claims 1 to 13.

Images