JP2001127977A - Device and method for correcting document image - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device and a method for automatically correcting a multilevel document image in a direction at high speed and for automatically composting the multilevel document image at the high speed, which is inputted by division plural number of times. SOLUTION: An inputted multilevel image (a color image and a gray scale image) is stored in an image memory and also processed by binarization. The a direction and a superimposing position are detected concerning binarized image data. A direction, detecting result and a superimposing position detecting result which are obtained concerning the binarized image, are used to execute the correct processing of the multilevel picture which is stored in the image memory and, then, it is adopted as an output image. Thus, the document image which is transmitted by an operation being different from the normal one in direction or the document image which transmitted by division into the plurality of image through the use of a hard-held scanner, etc., is stored as one kind of erect image data and are outputted.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus and a method for correcting a multi-valued image input from an image input device such as a scanner.

[0002]

2. Description of the Related Art In recent years, in addition to a stationary flat head scanner, a small, hand-operated, hand-held scanner that is easy to carry has been developed and commercialized.

[0003] Since the handheld scanner is a manual scan, the user can scan the document to be read from either the vertical or horizontal direction, but there is a problem that the image is rotated or the mirror image is inverted. In order to solve these problems, a technique for automatically correcting the orientation of an image has been developed (see Japanese Patent Application No. 10-146926).

[0004] Further, the hand-held scanner has a small body, so that the width at which scanning can be performed at one time is narrow. In order to capture a large image exceeding the width of the scanner, it is necessary to capture the image in a plurality of times and combine the images. A technique has been developed in which an image is captured in a plurality of times so as to form an overlapping area, and when the image is captured in the buffer memory, the image is automatically combined and the resultant image is recorded in the buffer memory ( See Japanese Patent Application No. 11-111708).

[0005] These techniques are for correcting a binary-input document image, and no technique for correcting a multi-level input document image has been developed.

Here, the multi-valued image is an image in which pixels constituting the image are represented by multi-valued parameters or a plurality of multi-valued parameters. For example, a gray-scale image is a multi-valued image because the only parameter that expresses color is luminance, but the possible values of luminance are a plurality of values. Another example is a color image. A color image, for example,
It is a multi-valued image in the sense that one pixel is represented by three parameters of R value, G value and B value, and each parameter can take a plurality of values.

[0007]

Conventional techniques for automatically correcting the direction of an input document image and techniques for automatically synthesizing a divided and input document image automatically correct a binary document image. A technique for correcting a multi-valued document image has not been developed.

Also, since a multi-valued image has a larger data amount than a binary image, processing the multi-valued image data itself requires a larger amount of memory and a longer processing time. Problems arise.

An object of the present invention is to provide an apparatus and a method for automatically and rapidly correcting a direction of a multi-valued document image and further automatically and rapidly synthesizing a multi-valued document image inputted in a plurality of times. Is to provide a way.

[0010]

An apparatus according to the present invention comprises a binarizing unit for binarizing acquired image data, and correcting information from the binary image data binarized by the binarizing unit. A correction information generating unit configured to generate the correction information, and a correction unit configured to correct the image data of the multi-valued image obtained by the image obtaining unit based on the correction information obtained by the correction information generating unit. .

The method of the present invention comprises the steps of: (a) binarizing the acquired image data; and (b) the step (a).
Generating correction information from the binary image data binarized in step (c), and (c) an image of a multi-valued image obtained by the image obtaining means based on the correction information obtained in step (b) Correcting the data.

According to the present invention, a binary image is generated from a multi-valued image, correction information is generated based on the binary image, and the multi-valued image is corrected using the correction information. Therefore, the correction of the multi-valued image can be performed without consuming a large amount of memory or taking a lot of processing time to perform the correction processing on the multi-valued image as it is.

In particular, when a multi-valued image is acquired by using a handheld scanner, even when the direction of the image is different from the normal direction or when one image is divided into a plurality of images, the multi-valued image is acquired. Since the image is not processed as it is, but is converted into a binary image and the correction information is obtained before processing, the memory capacity for the processing can be small, and the multi-valued image can be quickly converted to one.
It can be corrected to an erect image.

[0014]

FIG. 1 is a diagram for explaining the principle of the present invention.

The present invention is characterized in that the direction of a multi-valued document image is automatically corrected, and the multi-valued document images input in a plurality of times are automatically combined.

First, in step S1, an image is input. The image is read using a device capable of reading a multi-value image such as a color scanner. The input multi-valued image data is branched into two, and one is binarized in step S3. Then, in step S4, 2
The character direction and the overlapping position are detected from the digitized image.

The other of the branched multivalued image data is stored in a memory for recording an input image in step S2. When the direction detection and the overlap position detection processing in step S4 are completed, in step S5, the multi-valued image data is read from the input image recording memory,
Correction processing is performed based on the detection result obtained in step S4. When the correction processing in step S5 is completed, the multi-valued image that has been rotated, inverted, or divided into a plurality of pieces is configured as one erect document image, which is output in step S6. Output as an image.

With the above processing, the direction of a multi-valued image can be automatically corrected, and a plurality of document images can be automatically combined.

The direction detection in step S4 in FIG. 1 uses a "document image correction apparatus and method" described in Japanese Patent Application No. 10-240871. A “document image combining device, a document image combining method, and a document image combining program” described in Japanese Patent Application Laid-Open No. 11-00381 are used.

The binarization processing can be performed by a method of processing by focusing on the RGB values of the color components of the input image, or a method of converting the RGB values of the color components into luminance components and processing by focusing on the luminance components. .

The method of determining from the color components of the input image is R
Focusing on one component of the RGB values, the pixel whose component exceeds the threshold is set to white, and all the other pixels are set to black. The pixel in which at least two components of the RGB value exceed the threshold is set as the pixel. Two methods can be considered in which white is set and all other pixels are set to black. Focusing on one component of the RGB values is a case where the scanner does not have all the R, G, and B light sources, that is, a case where the scanner is a grayscale image acquisition scanner. . R
Attention is paid to two or more components of the GB value in the case of a full-color scanner. In the case of full color, it is known that in a general document, the background color is colored with a relatively light color, and the characters are colored with a relatively dark color. Of course, the opposite may be the case, such as white characters,
There is a tendency for the background color and the character color to be contrasting colors. In this case, generally, a bright color is one of the three components of the RGB values.
It is known that at least two components have a large value, and a binarization method that makes white when two or more components of RGB values are equal to or larger than a threshold value utilizes this fact. is there.

Also, a bright color tends to have a high luminance component, and a dark color tends to have a low luminance component. So, RGB
It is also possible to use a method in which the value is converted to a luminance component, and pixels in which the luminance component exceeds the threshold value are white, and all other pixels are black.

FIG. 2 and FIG. 3 are diagrams showing an example of how to obtain a threshold.

When the threshold value is obtained, a method of expressing the color components of the input image in a histogram for each color component and setting the center of two peaks formed on the histogram as a threshold value (FIG. 2)
The two typical methods are a method of calculating the maximum value and the minimum value of the RGB components, and using the center of the maximum value and the minimum value as a threshold (FIG. 3). FIGS. 2 and 3 each show one component of the RGB components in a graph. When a threshold value is obtained, processing is performed on all components.

In the method of FIG. 2, since a document image is largely divided into a base color and a character color, a histogram is obtained as shown in FIG.
There are two large mountains like. Therefore, attention has been paid to the fact that if the center of the two peaks is set as the threshold value, the background portion and the character portion can be divided into binary values.

In some cases, such as when a document image uses a plurality of background colors or a plurality of character colors, a plurality of peaks can be formed on the histogram, so that the threshold value cannot be determined by the former method. is there. In this case, by taking the middle between the maximum value and the minimum value of the RGB components as shown in FIG. 3, the background portion and the character portion can be divided into two values.

Therefore, in general, when determining a threshold value, a histogram of RGB components is created, the number of peaks is counted, and when only two peaks are present per component, the method of FIG. If there are three or more peaks, the method of FIG. 3 is used.

FIGS. 4 to 6 are diagrams for explaining the correction process in the character direction.

As shown in FIG. 4, when a horizontally written document is captured using a handheld scanner, image data is sent from the scanner as needed and temporarily stored in a buffer memory. When a certain amount of data has been sent, a part of the data stored in the buffer is extracted, and the extracted image is binarized. Then, as shown in FIG. 5, the binary image is processed to detect the character direction of the image. In the detection, for example, one character (“A” in FIG. 5) is extracted, and the extracted character is rotated by an appropriate angle to perform recognition. Assuming that the angle at which the character can be determined to be a predetermined character as a result of the recognition is 0 °, it can be determined that the character is oriented by the determined angle in the direction opposite to the direction in which the character is rotated. If the character direction cannot be detected, a part of the data is extracted again from a different area, and the same processing is performed.

By processing a part of an image by extracting a part of the image, processing a part of the image requires a smaller amount of memory and processing at a high speed than processing the entire image. .

FIG. 6 is a diagram for explaining a method of storing image data input by a scanner.

For example, it is assumed that the image correction apparatus according to the embodiment of the present invention is installed in a Microsoft Windows application. In this case, the data sent from the input device is stored in a Dib format. However, the image may be stored in another format such as the PBM format.

In the buffer memory, bit map data is stored for each line in the order of scanning. Here, in the embodiment of the present invention, by changing the order or direction in which the multi-valued image is stored in the buffer memory based on the detection result, the multi-valued document image is erected and stored in the buffer memory. To be stored.

FIGS. 7 and 8 are views for explaining the state of document conversion after the detection of the character direction.

As shown in FIG. 7, when the character direction of the input image is detected, the data stored in the buffer memory up to that time is corrected, stored in the output image memory, and sent from the scanner. The incoming data is stored in the output memory while being corrected. By extracting a part of the data sent from the scanner, detecting the character direction first, and determining the character direction, the image data is stored in the output image memory while sequentially converting the image data. An erect document image can be obtained at a higher speed than when processing is performed when all inputs are completed. The method of correcting the data sent from the scanner is basically the same for a color image and a binary image, and the order of bitmap data arrangement is from vertical to horizontal or from horizontal to vertical. Rearrange and save in output image memory.

In the Windows Dib format, an image cannot be saved unless the size of the input image is known. Therefore, as shown in FIG. 8, the character direction is detected at the stage of inputting the image, and the image data stored in the buffer memory is converted into the direction that has been established at once when the input of the image is completed. Is also good.

FIGS. 9 to 14 are diagrams for explaining the processing of image overlap correction and direction correction.

As shown in FIG. 9, in order to detect the overlapping position of the images, it is necessary to change the direction of the image to an erect state. The character direction of the input image is detected by the method described above, and 2
Correct the direction of the value image. Then, after the direction of the binary image is corrected, the overlapping position is detected by using the method described in Japanese Patent Application No. 11-00381 “Document image combining device, document image combining method and document image combining program”.

As shown in FIG. 10, two images input from the scanner are stored in a buffer memory. If the capture of the second image has been completed after detecting the overlap position, the portion of the data of the first image and the second image required for the composition is converted into the detected character direction. , And store them while compositing.

If the capture of the second image has not been completed, the portion of the data of the first image stored in the buffer memory that is necessary for the composition is converted into the detected character direction and composed. While storing. Also, of the data of the second image stored in the buffer memory up to that point, the part necessary for the composition is converted into the detected character direction,
Store while compositing. Further, the data sent from the scanner is stored while being converted in the same manner.

As a result, the amount of memory required for processing can be suppressed, and processing can be performed at high speed.

FIG. 11 is a diagram showing an example of an image reading method.

The scanner is scanned in a direction parallel to the line of the document. Since the document shown in FIG. 11 is written horizontally, the scanner is scanned in the horizontal direction. If the document is written vertically, the scanner is scanned in the vertical direction.

Attention is paid to the point where the characters in the overlapping area of the image coincide with each other, and the image is captured so as to have the overlapping area in order to detect the overlapping position of the image.

As long as the scanning direction of the scanner is parallel to the row, the scanning may be performed from either direction. However, it is assumed that the data is always captured so as to have an overlapping area.

FIG. 12 is a diagram for explaining an example of a method of detecting an overlapping position of images.

First, the first image is captured. At the stage where the data is being sent, a part of the image is extracted to create a binary image. The erect direction of the image is detected from the binary image, and the binary image is corrected in the detected direction.

Next, the second image is captured. At the stage where the data is being sent, a part of the image is extracted to create a binary image. The erect direction of the image is detected from the binary image,
The binary image is corrected in the detected direction.

If the computer has room for correction other than the process of receiving the data, the image is binarized and the image is corrected in the erecting direction when the data is being sent. If there is no room for the correction processing, the correction processing may be performed after the image capturing is completed. This is based on the fact that image data reception is performed in the form of interruption to a computer. While the interruption occurs, the receiving device receives the image data, but in many cases, the computer itself does nothing, so when this operation is not being performed, the image is corrected in the erect direction. If
Processing can be performed more efficiently.

Next, an overlap position is detected from the two binary images corrected in the erect direction. The overlapping position detection method
Here, the method of Japanese Patent Application No. 11-111708 is used, but another method may be used.

FIGS. 13 and 14 show the detected overlapping positions.
3 shows a method of combining two multi-valued images.

After the second image has been captured,
FIG. 13 shows a combining method when the overlapping position of the images is detected, and FIG. 14 shows a combining method when the overlapping position is detected before the capture of the second image is completed.

Since the overlapping position is detected from the image obtained by correcting the binary image in the erect direction, the coordinates are converted to the overlapping position before the correction in the erect direction to obtain the overlapping position of the two images.

In the method of FIG. 13, two images are stored in accordance with the detected overlapping position while correcting in the detected erect direction.

In the method of FIG. 14, since the capture of the second image has not been completed, the first data and the data transmitted so far are corrected in the detected erect direction while Two images are stored according to the detected overlapping position. The transmitted data is stored while being corrected in the same manner.

From the above processing, multi-valued images can be combined.

FIG. 15 is a diagram showing an example of the configuration of a processing device for correcting the direction of a multivalued image and combining the documents.

The input device 10 is an image capturing device such as a color scanner capable of capturing a multi-valued image. When multi-valued image data input from the input device 10 is stored in the image memory 13 and color image correction is performed, the image memory 13
The image is read from and processed. On the other hand, a color image (grayscale image) read by the input device 10
Is input to the binarization processing unit 11 and binarized,
Value image. The binary image is input to the direction detection / overlap position detection unit 12, and image information (erect direction,
(Overlapping position) is detected. Image information detected from the binary image is input to the color image correction processing unit 14, where the multi-value (color) image input from the image memory 13 undergoes direction conversion and overlap processing. The multi-valued image corrected in this way is output to the image output unit 15 as an output image.

FIG. 16 is a flowchart showing the overall processing of the embodiment of the present invention.

First, in step S101, an image is read from an input device. Next, the user inputs whether or not to combine images, and determines whether or not the combination mode has been set according to the input (step S102). If the result of the determination in step S102 is NO, direction correction for correcting the input image to an erect image is performed (step S103), and if the determination in step S102 is YES, 2
The first image is read (step S104). And
In step S105, the first image read in step S101 and the second image read in step S104 are combined to generate and output a combined image, and the process ends.

With the above operation, the direction of a multi-value input document image can be automatically corrected, and a plurality of document images can be automatically combined. Further, the amount of memory required for processing can be suppressed, and processing can be performed at high speed.

FIG. 17 is a flowchart showing an example of the document image direction correction processing in step S103 of FIG.

First, in step S101, the multivalued image data fetched from the input device is stored in the image memory. Next, in step S107, it is determined whether all image data has been processed. At the beginning of the process, since the process of the image data has just started, the process proceeds to step S102. In step S102, a part of the data stored in the image memory is extracted. Step S
At 103, some of the extracted data is binarized. In step S104, the erect direction of the image is detected from the binarized monochrome image. The detection of the erect direction uses the method described in the specification of Japanese Patent Application No. 11-00381.

Next, in step S105, it is determined whether or not the erect direction of the image has been detected. If the erect direction has not been detected, the process returns to step S107, and a data area different from the first data area or the first data area has The data area + another data area is taken out, and the orientation of the image in the erect direction is detected again. This process is repeated until the erect direction of the image is detected. In step S105, if the direction cannot be detected and the direction cannot be detected even when processing is performed on all regions of the image, the determination in step S107 is YES, and the process ends. Step S
If the erect direction of the image can be detected in 105, the process proceeds to step S106.

In step S106, step S104
The multi-valued color image stored in step S101 is corrected based on the erect direction of the image detected in.

With the above processing, a multi-valued input document image can be corrected in the erect direction.

FIG. 18 is a flowchart showing an example of the document image combining process in step S105 of FIG.

First, in step S201, the first image is captured. In step S202, the captured first image data is stored in the image memory. Then, in step S217, it is determined whether or not all the image data has been processed. If all of the image data has already been processed, the process ends because the direction cannot be detected. In step S217,
If the processing has not been performed for all the image data, a part of the stored image data is extracted in step S203. Then, in step S204,
The extracted part of the image data is binarized. Then, in step S205, the erect direction of the image is detected from the binarized monochrome image. In step S206, it is determined whether the erect direction of the image has been detected. If the erect direction has not been detected, the process returns to step S217. If it is determined in step S206 that the direction has been detected, the binary image is corrected in the erect direction in step S207, and the process proceeds to step S215.

Next, after all the first images have been captured, the second image is captured in step S208. In step S209, the captured second image data is stored in the image memory. Next, in step S218, it is determined whether or not all image data has been processed. If processing has been completed for all image data, it is determined that the erect direction cannot be detected for the second image. The process ends. If it is determined in step S218 that all image data has not been processed, the process proceeds to step S210, where a part of the image data stored in the image memory is extracted, and in step S211 the binarization is performed. I do. Then, step S2
At 12, the direction is detected, and at step S213, it is determined whether or not the direction detection is successful. If the direction detection has failed, the process returns to step S218, and the processes from step S210 to step S213 are repeated. If it is determined in step S213 that the direction detection has been successful, the binary image is corrected in the erect direction in step S214. Then, step S207
And 2 corrected in the erect direction obtained in step S214.
In step S215, the overlapping position is detected using the value image. Then, in step S219, it is determined whether or not the overlap position has been detected. If the overlap position has not been detected, the process ends. In step S219,
If the overlap position has been detected, the process proceeds to step S216.

The first image data stored in step S202, the erect direction of the first image detected in step S205, the second image data stored in step S209, and the detection in step S212 The erect direction of the second image thus obtained, and the overlapping position of the two images detected in step S215 are used for the correction processing in step S216. In step S216, the two input images are combined at the detected overlapping position while correcting the detected images in the erect direction.

By the above processing, two document images input in multi-value can be automatically combined.

FIG. 19 is a diagram showing a hardware environment for executing a program when the embodiment of the present invention is realized by a program.

When the power is turned on, the CPU 21 reads the BIOS and the like from the ROM 22 via the bus 20, and controls the input / output device 30, the communication interface 24, the recording medium reader 28, the storage device 27, and the RAM 23.

When the computer 31 is used exclusively for realizing the embodiment of the present invention, the program for realizing the present embodiment is stored in the ROM 22 so that the CPU 21 reads and executes the program. I do.

Alternatively, the program may be recorded in a storage device 27 such as a hard disk, expanded as necessary in the RAM 23, and executed by the CPU 21. Also, if the program is a floppy disk,
When the program is distributed on a portable recording medium 29 such as a CD-ROM or a DVD, the user purchases the program recorded on the portable recording medium 29, reads the program with the recording medium reading device 28, and stores the program in the storage device 27. It may be stored and used. Alternatively, R
The CPU 21 may sequentially execute the program while reading the program into the AM 23.

In implementing the embodiment of the present invention, the input / output device 30 includes a scanner, which is an image reading device, and an image output device (such as a printer) in addition to a normal keyboard, mouse, and display. Need to be.
However, the computer 31 is dedicated to image acquisition, or
In the case of exclusive use for image output, a configuration including only one of the scanner and the image output device may be employed. The image data read from the scanner of the input / output device 30 is
Portable recording medium 29, storage device 27, or ROM2
The program is processed by executing the program recorded in the input / output device 2 and is output from an image output device provided in the input / output device 30. Further, the image data captured by the scanner may be transmitted to the network 25 via the communication interface 24, and the image data may be transmitted to another user connected via the network 25. The other user has a device equivalent to the computer 31, processes the transmitted image data, and causes the image output device to output an image.

A program for implementing the embodiment of the present invention can be downloaded from an information provider 26 via a network 25 connected to the communication interface 24 and acquired. The downloaded program is stored in the storage device 27 or the like, and the CPU 21 executes the program expanded in the RAM 23 as necessary, whereby the embodiment of the present invention can be realized. When the network 25 is a LAN or the like, the program can be executed in a network environment without downloading.

[0078]

According to the present invention, a multi-valued document image input from either the vertical or horizontal direction using a hand-held scanner can be automatically corrected in the erect direction. Further, even if the document exceeds the width of the scanner, the image can be automatically combined into one document image by inputting the image in a plurality of times.

Further, in order to binarize the input multi-valued image and detect the document direction and the overlapping position from the binarized image,
The amount of memory required for processing can be reduced and processing can be performed at high speed.

Therefore, the present invention greatly contributes to improvement of operability of image input by the handheld scanner and improvement of the user interface.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is a diagram (part 1) illustrating an example of how to determine a threshold.

FIG. 3 is a diagram (part 2) illustrating an example of how to obtain a threshold.

FIG. 4 is a diagram (part 1) for explaining a character direction correction process;

FIG. 5 is a diagram (part 2) for explaining a character direction correction process;

FIG. 6 is a diagram (part 3) for explaining a character direction correction process;

FIG. 7 is a diagram (part 1) illustrating a state of document conversion after character direction detection.

FIG. 8 is a diagram (part 2) for explaining the state of document conversion after character direction detection.

FIG. 9 is a diagram for explaining processing of image overlap correction and direction correction.

FIG. 10 is a diagram (part 1) for explaining processing of image overlap correction and direction correction;

FIG. 11 is a diagram (part 2) for explaining processing of image overlap correction and direction correction;

FIG. 12 is a diagram (part 3) for explaining processing of image overlap correction and direction correction;

FIG. 13 is a diagram (part 4) for explaining the process of image overlap correction and direction correction;

FIG. 14 is a diagram (No. 5) for explaining the processing of image overlap correction and direction correction;

FIG. 15 is a diagram illustrating a configuration example of a processing device that performs direction correction and document combination of a multivalued image.

FIG. 16 is a flowchart illustrating an overall process according to the embodiment of the present invention.

17 is a flowchart illustrating an example of a document image direction correction process in step S103 of FIG.

FIG. 18 is a flowchart illustrating an example of a document image combining process in step S105 of FIG.

FIG. 19 is a diagram illustrating a hardware environment for executing a program when the embodiment of the present invention is implemented by the program.

[Explanation of symbols]

 Reference Signs List 10 input device 11 binarization processing unit 12 direction detection / overlapping position detection unit 13 image memory 14 color image correction processing unit 15 image output unit

Continuation of the front page (72) Inventor Tsuguo Noda 4-1-1, Kamidadanaka, Nakahara-ku, Kawasaki-shi, Kanagawa F-term within Fujitsu Limited (Reference) 5B057 CA01 CA08 CA12 CB01 CB06 CB08 CB12 CD03 CE09 CE10 CH01 CH11 DA07 DB02 DB06 DB09 5C076 AA11 AA19 AA24 AA36 BA03 BA06 5C077 LL18 MM18 MP08 NP01 PP22 PP23 PP32 PQ12 PQ22 RR02

Claims (9)

[Claims] 1. Binarization means for binarizing acquired image data, correction information generation means for generating correction information from binary image data binarized by the binarization means, A correction unit configured to correct the image data of the multi-valued image obtained by the image obtaining unit based on the correction information obtained by the information generation unit. 2. The image forming apparatus according to claim 1, wherein the correction information is direction information indicating a direction of the image data with respect to an erect state, and the correction unit corrects an image direction. The apparatus of claim 1, wherein: 3. The image forming apparatus according to claim 1, wherein the correction information is overlapping position information indicating an overlapping state between the plurality of image data when one image is divided into a plurality of image data. The apparatus according to claim 1, wherein: 4. The binarizing means generates the binary image data by determining whether at least one of the color components of the image data exceeds a predetermined threshold. The device of claim 1, wherein 5. The binarizing means generates the binary image data by determining whether at least two or more of the color components of the image data exceed a predetermined threshold value. The device of claim 1, wherein 6. The binarized image data by extracting a luminance component of the image data and determining whether or not the luminance component of the image data is larger than a predetermined threshold value. The apparatus according to claim 1, wherein the apparatus generates: 7. The apparatus according to claim 1, wherein said correction information generating means generates correction information for binary image data corresponding to a part of said image data, and corrects the entire image data using said correction information. The apparatus of claim 1, wherein: 8. A step of: (a) binarizing the acquired image data; (b) generating correction information from the binary image data binarized in the step (a); Correcting the image data of the multi-valued image obtained by the image obtaining means based on the correction information obtained in the step (b). 9. A computer comprising: (a) binarizing acquired image data; and (b) generating correction information from the binary image data binarized in step (a). And (c) correcting the image data of the multi-valued image obtained by the image obtaining means based on the correction information obtained in the step (b). A computer-readable recording medium on which a program is recorded.