JP3106080B2 - Image processing apparatus and method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus and method, for example, by determining a binarization threshold value of a multi-valued image.
The present invention relates to an image processing device and a method for performing binarization.

[0002]

2. Description of the Related Art In recent years, image processing techniques have been remarkably developed, and image processing apparatuses capable of processing multi-valued images such as full-color images and character recognition processing in multi-valued images have become widespread.

In such an image processing technique, binarization processing of a multi-valued image is an indispensable technique. Conventional 2
Examples of the binarization method include a simple binarization method using a preset fixed threshold, and a threshold at which the inter-class variance is maximized when the histogram is divided into two classes at a certain threshold. Otsu method (Otsu, "How to automatically select thresholds based on discrimination and least squares criterion", IEICE Transactions, vol.J63-D, no.4, pp.349-356, 1980), Further, there has been a binarization method for setting a threshold value for an image having a gradation according to a local density.

[0004]

However, the above-described conventional binarizing method in the image processing apparatus has the following problems.

In the simple binarization method using a fixed threshold value, it is difficult to set an appropriate threshold value between the object density and the background density in an image. As a result, the entire image may be crushed black, or conversely. It had turned white. Further, according to the Otsu method, when the distributions of the two classes are extremely different, the threshold value is closer to the larger class, so that a binary image with much noise has been generated. Further, in the binarization method in which the threshold is set according to the local density, since the image is locally divided, there is a problem that block distortion is likely to occur.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problem, and an image processing apparatus and an image processing apparatus capable of automatically setting an appropriate binarization threshold value between an object density and a background density in an image. It is intended to provide such a method.

[0007]

In order to achieve the above-mentioned object, the present invention has the following arrangement.

That is, a calculating means for calculating a luminance frequency distribution of a multi-valued image, and a threshold value determining for determining a binarization threshold for separating a background and an object in the multi-valued image based on the luminance frequency distribution Means, wherein the threshold value determining means sets first and second luminance values as a start point and an end point of the specific luminance area in the luminance frequency distribution, respectively, and An average brightness value and distribution bias are obtained, and when the distribution bias does not satisfy a predetermined condition, the specific brightness area is newly set by changing either the first brightness value or the second brightness value. The process of newly calculating the average luminance value and the distribution deviation in the specific luminance region is repeated until the distribution deviation satisfies the predetermined condition, and when the distribution deviation satisfies the predetermined condition, An average luminance value in the specific luminance area, and sets a binarization threshold.

[0009] Further, it is characterized in that it comprises an input means for inputting the multi-valued image, and a binarizing means for binarizing the multi-valued image using the binarization threshold.

For example, the threshold value determining means calculates the distribution bias based on a difference between a luminance value of each pixel in the specific luminance area and an average luminance value in the specific luminance area in the luminance frequency distribution. It is characterized by the following.

For example, the pre-threshold determining means determines the distribution bias based on an odd power of a difference between a luminance value of each pixel in the specific luminance region and an average luminance value in the specific luminance region in the luminance frequency distribution. Is calculated.

For example, the threshold value determining means determines that the predetermined condition is satisfied if the distribution bias is within a predetermined range, and if the distribution bias is outside the predetermined range and is positive,
The first luminance value is changed to an average luminance value in the specific luminance region, and if the distribution bias is outside the predetermined range and negative, the second luminance value is changed to an average luminance value in the specific luminance region. It is characterized by changing.

[0013]

According to the above arrangement, in the input multi-valued image, based on the luminance frequency and its bias, the region where the most suitable threshold value for separating the background and the object in the image is specified, The specific operation and effect that the binarization threshold can be determined based on the average luminance value of the specific region is obtained.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment according to the present invention will be described below in detail with reference to the drawings.

<First Embodiment> FIG. 1 is a block diagram showing a system configuration for executing a binarization process in this embodiment. In FIG. 1, reference numeral 1 denotes an image processing device for performing a character recognition process, 2 denotes an image input device such as a scanner for inputting an image, and 3 denotes an image display device for displaying a processed image.

In the image processing apparatus 1, 4 is an input section serving as an interface with the image input apparatus 2, 5 is a storage section such as a memory for storing data being processed, and 6 is a luminance frequency (histogram) of the input image. This is the luminance frequency accumulating unit that performs the operation. Reference numeral 7 denotes a binarization threshold calculator that calculates a binarization threshold of the input image, and 8 denotes a binarization unit that creates a binary image using the threshold calculated by the binarization threshold calculator 7. . 9
Is an area separation unit that separates an image into areas for each attribute.
0 is a character recognition unit that performs character recognition processing on an area extracted as a character area by area separation, 11 is an image processing unit that performs various types of image processing on an area separated from the character area, and 12 is an image processing unit 3 This is an output unit that serves as an interface. These components are generally controlled by a CPU (not shown).

The OCR processing executed in the image processing apparatus 1 according to the present embodiment having the above-described configuration will be described below.

FIG. 2 is a flowchart showing an image area separation OCR process using a binarization threshold value determination method which is a feature of the present embodiment.

First, in step S501, image data is input by the image input device 2 such as a scanner. The input here is performed as 8-bit multivalued image data. Subsequently, in step S502, step S5
For the multi-valued image input at 01, a binarization threshold optimal for image area separation, which is a feature of the present embodiment, is determined, and a binary image is generated using the binarization threshold. In step S503, image area separation of the binary image generated in step S502 is performed, and area data to which the attribute is added is output. Next, in step S504, step S504
The erroneous determination result included in the area data separated in 503 is removed. Hereinafter, the erroneous determination result removal processing in step S504 is referred to as layout noise reduction (LNR).
Called. Then, the process proceeds to step S505, and step S505
In the area data separated in 503, an area designated as “text” is cut out from the binary image.
OCR processing is performed on the value image, and the recognized character code is output.

<< Binarization Processing >> Next, the binarization processing in this embodiment will be described in detail with reference to the flowchart of FIG.

In FIG. 3, first, in step S1, an 8-bit multivalued image is input from a storage unit 5 in the image processing apparatus 1 to a memory (not shown). It is assumed that the multivalued image is stored in the storage unit 5 in advance by the image input device 2 such as a scanner. Then, in step S2, a histogram of the entire input image is calculated. Here, all the pixels in the image are used to calculate the frequency for 8 bits, that is, each digital value from “0” to “255”. Thereby, for example, a histogram shown in FIG. 4 is obtained.

Next, in step S3, the parameter S
"0" and "255" are set in TART and END, respectively. START and END correspond to the start point and the end point of the statistic of the luminance value obtained in the subsequent steps S4 and S5, respectively.

In step S4, START to END
The average value AV of the pixels corresponding to the digital values up to is calculated. For example, if START = 0 and END = 255, an average value AV of pixels having values from “0” to “255” (all pixels in this case) is calculated, and START = 0, END =
If it is 177, the average value AV of pixels having values from “0” to “177” is calculated.

In step S5, START to END
The skew value SK of the pixel corresponding to the luminance value up to is calculated. The skew value is a statistic indicating the bias of the histogram distribution. To calculate the skew value, the following (1)
Use the formula.

SK = (Σ (Xi-AV) ∧3) / D
(1) (Note that the notation R∧3 indicates the cube of R.) Here, Xi is the luminance value of the pixel. D is a variance value of the entire image, and is calculated by equation (2).

D = Σ (Xi−AV) ∧2 (2) (R∧2 represents the square of R.) In the above equation (1), the skew value is It is calculated by raising the difference between the luminance value of each pixel and its average value to the third power, but is not limited to the third power if it is an odd power.

Subsequently, in steps S6 and S7, the direction of the bias of the histogram is determined. First, in step S6,
The bias direction of the histogram is determined by the following equation (3). This is to determine whether the bias of the histogram is in a range of values smaller than the average value AV.

SK <−1.0 (3) If Expression (3) is true in Step S6, Step S1
The process proceeds to 0, and if false, the process proceeds to step S7. Step S1
At 0, START is not changed, and the average value AV is added to END.
Is set. Then, returning to step S4, the STA
The average value AV from the RT value to the END value is calculated.

On the other hand, in step S7, the following (4)
The direction of the bias of the histogram is determined by the equation. This is to determine whether the bias of the histogram is in a range of values larger than the average value AV.

SK> 1.0 (4) If equation (4) is true in step S7, step S1
The process proceeds to step S1, and if false, the process proceeds to step S8. Step S1
At 1, the average value AV is set in START, and END is not changed. Then, returning to step S4, the STA
The average value AV from the RT value to the END value is calculated.

On the other hand, in step S8, steps S6, S
The average value AV when the conditions in 7 are both false is 2
The threshold value TH is set. Then, in step S9, a simple binarization process using the binarization threshold TH is performed.

As described above, the binarization processing in the present embodiment is performed, but the range shown by equations (3) and (4) is not limited to this.

Hereinafter, the binarization processing of this embodiment will be described in more detail with reference to specific examples of images.

Using the example of the histogram shown in FIG. 4 described above, the process of determining the binarization threshold TH in this embodiment will be described.

FIG. 4 shows a histogram of a certain image (8-bit input). In FIG. 4, the horizontal axis represents the digital value of the luminance representing "0" or black at the left end and "255" or white at the right end, and the vertical axis represents the frequency of each digital value. FIG. 5 is a diagram showing a change in the value of each parameter during the processing shown in steps S4 and S5 in the above-described binarization processing shown in FIG. 3 for an image having a histogram as shown in FIG. It is. Each parameter value shown in FIG. 5 is indicated by the number of times of passing through steps S4 and S5.

First, 1 which passes through steps S4 and S5
In the third process, the average value AV and the statistic SK are calculated with START = 0 and END = 255, and “177”,
The value “−78.9” is obtained. In this case, the statistic SK
Is less than “−1.0”, START = 0 and END = 177 are set in step S10.

Subsequently, in the second processing, START = 0,
The average value AV and the statistic SK at END = 177 are calculated to obtain values “91” and “−8.6”, respectively.
Again, since the statistic SK is less than "-1.0", START = 0 and END = 91 are set in step S10.

Subsequently, in the third processing, START =
The average value AV and the statistic SK at 0, END = 91 are calculated to obtain values of "43" and "9.6", respectively.
In this case, since the statistic SK exceeds “1.0”, START = 43 and END = 91 in step S11.
Is set.

Subsequently, in the fourth processing, START = 4
3, the average value AV and the statistic SK at END = 91 are calculated to obtain values "72" and "-7.0", respectively. Again, since the statistic SK is less than “−1.0”, START = 43, END in step S10.
= 72 is set.

Subsequently, in the fifth processing, START = 4
3, the average value AV and the statistic SK at END = 72 are calculated to obtain values of “58” and “−2.2”, respectively. Again, since the statistic SK is less than “−1.0”, START = 43, END in step S10.
= 58 is set.

In the sixth processing, START = 4
3, the average value AV and the statistic SK at END = 58 are calculated to obtain values “50” and “−0.4”, respectively. Here, the statistic SK is “−1.0” or more and “1.
0 "or less, the conditions of Steps S6 and S7 are not satisfied, and the process proceeds to Step S8 to set the binarization threshold TH as
“50” is set. Then, in step S9,
Simple binarization processing using the binarization threshold TH is performed, and the binarized image is stored in the storage unit 5.

As described above, in this embodiment, the binarization threshold is determined so that the skew value converges to a predetermined value, and binarization is performed. That is, in the input multi-valued image, based on the luminance frequency and its bias, after specifying the region where the threshold value most suitable for separating the background and the object in the image exists, the average of the specific region is determined. The luminance value is used as a binarization threshold. This makes it possible to automatically determine the optimum threshold value when classifying the luminance value of each pixel in the area on the multi-value input image into two classes, the background and the object.

<< Image Area Separation Processing >> Next, FIG.
The image area separation process shown in step S503 will be described in detail with reference to the flowchart in FIG.

First, in step S601, a binary image is input and stored in the storage unit 5. And step S6
In 02, the input image is thinned out so that m × n pixels become one pixel, and an image for image area separation is generated. At this time, if at least one black pixel exists in the m × n pixels, the m × n pixel is set as one black pixel.

Then, in step S603, for all the pixels of the image area separation image, the area where black pixels are continuous by a predetermined number in the up, down, left, right, and oblique directions is defined as one area. At this time, labels are assigned to the respective regions by assigning numbers to the regions in the order of detection. Next, in step S604, the regions are classified according to the width, height, area, and black pixel density of each region, and attribute labeling is performed. As the attribute of the area, details will be described later,
There are "table", "outer frame area", "text" and the like.

In step S605, the average of the widths and heights of all the regions labeled "text" is calculated, and if the obtained average width is larger than the average height, it is determined that the processed image is horizontal. In the opposite case, the character set is determined by regarding vertical writing. At the same time, the average height is set for horizontal writing, and the average width is set for vertical writing to make the size of one character.

Further, from the histograms of all the "text" regions in the vertical direction (at the time of horizontal writing) or the horizontal direction (at the time of vertical writing) on the image for image area separation, columns of sentences and line intervals are detected. In step S606, an area having a large character size in the “text” area is set to “title”. Then, in step S607, the "title" area and the "text"
The regions are merged according to the distance from the surrounding regions to form one integrated region. Next, in step S608,
It outputs area data such as attributes and coordinates and size in the original image for each area.

By performing the above processing, in this embodiment, image area separation processing of a binary image is performed, and each area data is obtained.

FIG. 7 shows an example of the above-mentioned area data.
Each area data item shown in FIG. 7 will be described below. "No.": Indicates the detection order of the areas. "Attribute": Indicates attribute information of an area, and eight types shown below are prepared.

"Route" Indicates that the image is the input image itself.

"Text" Indicates a character area.

"Title" Indicates a title area.

"Table" Indicates a table area.

"Noise" indicates a region that could not be determined as either a character region or an image region.

[Outer frame] Indicates an area such as a ruled line.

"Photo image" Indicates a photo area.

"Line image" Indicates a line image area. "Start point coordinates": X and Y coordinates of the area start in the original image. "End point coordinates": X and Y coordinates of the end of the area in the original image. "Number of pixels": Indicates the total number of pixels in the area. "Character set information": Indicates three types of character set information: vertical writing, horizontal writing, and unknown.

For the area data shown in FIG.
Only the area indicated by “text” hierarchically holds the area data (row area data) related to the rows before the merging in step S607 shown in FIG.

As described above, in this embodiment, the image area separation processing is performed. Note that the area data shown in FIG. 7 is merely an example to which the present embodiment is applied, and for example, other information may be appropriately added or reduced depending on the image processing apparatus.

<< LNR Processing >> Next, the LNR processing shown in step S504 of FIG. 3 will be described in detail with reference to the flowchart of FIG. The LNR process is a process of removing an image area separation error area from each of the image area separated areas.

First, in step S701 in FIG. 8, it is determined whether or not each area data after image area separation is a root area. The root area is an area surrounding the entire image, that is, the entire area. If the area is the root area, the process proceeds to step S706, and the LNR process is not performed. If it is not the root area, the process proceeds to step S702, and it is determined whether the area is a text area (character area) or a noise area. If it is a text area or a noise area, the process proceeds to step S703; otherwise, the process proceeds to step S705.

In step S703, the LNR in which the area data is removed as an area separation error according to the area size
Processing 1 is performed, and then, in step S704, LNR processing 3 is performed in which the area data is removed as an area separation error according to the black ratio in the area. On the other hand, in step S705, LNR processing 2 is performed in which region data that is neither a text region nor a noise region is removed as a region separation error according to the size of the region. Steps S703 and S703
The LNR processes 1, 3, and 2 in 704 and S705 will be described in detail below.

Then, in step S706, it is determined whether or not the processing for all the areas has been completed. If the processing has not been completed, the process returns to step S701; if completed, the LNR processing ends.

Hereinafter, first, the LNR shown in step S703 will be described.
Processing 1 will be described in detail.

FIG. 9 is a flowchart showing LNR processing 1. First, in step S731, the height H1 and the width W1 are referred to from the area data of the processing target area. Then, in order to calculate the height threshold HT1 and the width threshold WT1 used for determining the size of the region, in step S732,
A reading resolution SR of the image input device 2 such as a scanner;
The minimum number of points MP1 of the characters not to be removed in the image is set as MP1h and MP1w for the height and width, respectively.

In this embodiment, the threshold values HT1 and WT1 are calculated by the following equations (5) and (6).

HT1 = (SR / 72.0) × MP1h (5) WT1 = (SR / 72.0) × MP1w (6) In step S733, the threshold value of the height is calculated by the equation (5). HT
1 is calculated. For example, when the resolution SR of the image input device 2 is 400 dpi and the number of height points M of the minimum characters in the image is M
If P1h is 4 points, the height threshold HT1 is calculated as “22”. Then, step S734
Then, the height H1 of the area data is compared with the height threshold value HT1 calculated in step S733. If the height H1 of the area data is larger than the threshold HT1, step S735
Proceeds to step S738 if it is smaller than the threshold HT1.
Proceed to.

In step S735, the width threshold value WT1 is calculated by equation (6). Subsequently, in step S736,
The width W1 of the area data is compared with the width threshold WT1 calculated in step S735. If the width W1 of the area data is larger than the threshold WT1, the LNR process 1 ends. On the other hand, if the width W1 of the area data is smaller than the threshold WT1, the process proceeds to step S737, and the height H of the area data
The ratio H1 / W1 between 1 and the width W1 is determined. If this ratio is equal to or less than “2”, the LNR process 1 ends.
On the other hand, if the ratio exceeds “2”, it is determined that the area being processed has an area separation error, and the process advances to step S738 to remove the area.

Next, the LN shown in step S704 of FIG.
The R processing 3 will be described in detail with reference to the flowchart in FIG. First, in step S741, the number of black pixels BC in the area is accumulated. And step S742
Then, the black ratio BR in the area is calculated by the following equation (7).

BR1 = BC / (W1 × H1) × 100 (7) Next, in step S743, the minimum black ratio BRT1
And the maximum black ratio BRT2. BRT1 and BR
T2 is preset according to the black ratio characteristics of the character,
For example, BRT1 = 5, BRT2 = 52.

In step S744, the black ratio B in the area
R is compared with the minimum black ratio BRT1 and the maximum black ratio BRT2. If the black ratio BR is smaller than the minimum black ratio BRT1 or larger than the maximum black ratio BRT2, it is determined that the area being processed has an area separation error, and the process advances to step S745 to remove the area. Otherwise, the LNR processing 3 ends.

Next, the LN shown in step S705 of FIG.
The R processing 2 will be described in detail with reference to the flowchart in FIG. First, in step S751, the height H2 and the width W2 are referred to from the area data of the processing target area. Then, a height threshold value H used to determine the size of the region
Step S7 to calculate the threshold value WT2 of T2 and width.
At 52, the resolution SR of the image input device 2 is set. Then, in step S753, it is determined whether or not the attribute of the area being processed is an outer frame area. And
If it is an outer frame area, the process proceeds to step S754.

In step S754, the minimum number of points MP21 is set to MP21h,
Set as MP21w. Similarly, in step S757, the minimum point number MP22 is set as the height and width as MP22h and MP22w, respectively. Here, the minimum point numbers MP21 and MP22 represent the minimum size of the area not to be removed in terms of the character points, depending on whether or not the area is the outer frame area in the LNR processing 2.

Then, steps S755 and S758
, The height thresholds HT21 and HT22 and the width thresholds WT21 and WT22 are calculated by the above-described equations (5) and (6).
Is calculated. For example, if the resolution SR of the image input device 2 is 4
When the minimum point number MP22 is 4 points in both height and width at 00 dpi, the thresholds HT22 and WT22 are calculated as “22”. Then, Step S756
In step S759, a height threshold HT2 and a width threshold WT2 are set.

Subsequently, in step S760, the height H2 of the area data is compared with the height threshold HT2 set in steps S756 and S759, and the width W of the area data is compared.
The comparison with the width threshold value WT2 is performed in the same manner as in the case of FIG. If the height H2 of the area data is smaller than the threshold HT2 or the width W2 is smaller than the threshold WT2, it is determined that the area being processed has an area separation error, and the area is removed in step S761. Otherwise, the LNR process 2 ends.

As described above, the LNR processing of this embodiment removes an area determined to be an area separation error by three types of processing.

As described above, according to this embodiment, the optimum threshold value for classifying the luminance value of each pixel in the area on the multi-valued input image into two classes, the background and the object, is automatically determined. Can be determined. Therefore, the background and the object in the multi-valued image can be appropriately separated, and a highly detailed OCR process can be performed.

<Other Embodiments> The image input in the first embodiment has been described as 8-bit multi-valued image data. However, the present invention is not limited to this. And so on, that is, it is sufficient that there be information of a plurality of bits as image information for binarization.

The sampling of the image at the time of calculating the histogram is not limited to every pixel or every several pixels. Further, the calculation of the average AV, the statistic SK, and the like need not always be performed with eight bits, and may be performed with a small number of bits for speeding up, reducing memory, and the like.

Although the convergence condition of the skew value SK, which is a statistic, is ± 1.0, the invention is not limited to this. It is sufficient that the threshold value for binarization is determined using the skew value SK.

The present invention may be applied to a system including a plurality of devices such as an image scanner, a printer controller, and a printer, or may be applied to an apparatus including a single device such as a color copier. . Further, the present invention is not limited to the case where the image processing apparatus is provided with hardware as described above, and can be applied to a case where the present invention is achieved by supplying a program stored in a medium such as a magnetic disk to a system or an apparatus. Needless to say.

[0082]

As described above, according to the present invention, a luminance region having a threshold suitable for separating a background and an object in an image is specified based on the luminance frequency and its bias. Since the average luminance value in the specified luminance area is set as the binarization threshold value, an optimal threshold value for separating the background and the object can be automatically set regardless of the type of the input image.

[0083]

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a system configuration of an image processing apparatus according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating an image area separation OCR process according to the present embodiment.

FIG. 3 is a flowchart illustrating a binarization process in the embodiment.

FIG. 4 is a diagram illustrating an example of an image histogram according to the present embodiment.

FIG. 5 is a diagram illustrating an example of transition of each variable value in the binarization processing according to the embodiment.

FIG. 6 is a flowchart illustrating an image area separation process according to the present embodiment.

FIG. 7 is a diagram illustrating an example of area data according to the present embodiment.

FIG. 8 is a flowchart illustrating an area removal (LNR) process according to the present embodiment.

FIG. 9 is a flowchart showing an area removing process 1 based on the size of the area in the embodiment.

FIG. 10 is a flowchart illustrating a region removal process 2 based on a black ratio in the present embodiment.

FIG. 11 is a flowchart illustrating an area removal process 3 based on the size of the area in the present embodiment.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 image processing device 2 image input device 3 image display device 4 input unit 5 storage unit 6 luminance frequency accumulating unit 7 binarization threshold calculation unit 8 binarization unit 9 image area separation unit 10 character recognition unit 11 image processing unit 12 output Department

Continuation of the front page (56) References JP-A-61-140276 (JP, A) JP-A-5-191649 (JP, A) JP-A 8-223409 (JP, A) (58) Fields investigated (Int) .Cl. ⁷ , DB name) G06T 5/00 200

Claims (10)

(57) [Claims] 1. A luminance frequency distribution of a multi-valued image is calculated, and first and second luminance values are respectively set as a start point and an end point of a specific luminance area in the luminance frequency distribution, and an average luminance in the specific luminance area is set. Calculating a value and a distribution bias, when the distribution bias does not satisfy a predetermined condition, changing the first luminance value or the second luminance value to newly set the specific luminance region, The process of newly calculating the average luminance value and the distribution deviation in the specific luminance region is repeated until the distribution deviation satisfies the predetermined condition. When the distribution deviation satisfies the predetermined condition, the average in the specific luminance region is obtained. An image processing method comprising: setting a luminance value as a binarization threshold for separating a background and a target in the multi-valued image. 2. The image processing method according to claim 1, further comprising binarizing the multi-valued image using the binarization threshold. 3. The method according to claim 2, wherein the distribution bias is calculated based on a difference between a luminance value of each pixel in the specific luminance region and an average luminance value in the specific luminance region in the luminance frequency distribution. The image processing method according to claim 1. 4. The method according to claim 1, wherein the distribution bias is calculated based on an odd power of a difference between a luminance value of each pixel in the specific luminance region and an average luminance value in the specific luminance region in the luminance frequency distribution. The image processing method according to claim 3, wherein: 5. If the distribution deviation is within a predetermined range, it is determined that the predetermined condition is satisfied. If the distribution deviation is outside the predetermined range and is positive, the first luminance value is set to the specific luminance. 2. The method according to claim 1, wherein the second brightness value is changed to an average brightness value in the specific brightness area when the distribution bias is outside the predetermined range and is negative. The image processing method described in the above. 6. A calculating means for calculating a luminance frequency distribution of a multi-valued image, and a threshold value determining a binarization threshold value for separating a background and a target in the multi-valued image based on the luminance frequency distribution. Means, wherein the threshold value determining means sets first and second luminance values as a start point and an end point of the specific luminance area in the luminance frequency distribution, respectively, Calculating an average luminance value and a distribution bias, and when the distribution bias does not satisfy a predetermined condition, changing either the first luminance value or the second luminance value to newly set the specific luminance region Then, a process of newly obtaining the average luminance value and the distribution bias in the specific luminance region is repeated until the distribution bias satisfies the predetermined condition. When the distribution bias satisfies the predetermined condition, An image processing apparatus, wherein an average luminance value in a specific luminance area is set as a binarization threshold. 7. The image processing apparatus according to claim 6, further comprising: input means for inputting the multi-valued image; and binarizing means for binarizing the multi-valued image using the binarization threshold. The image processing apparatus according to any one of the preceding claims. 8. The method according to claim 1, wherein the threshold value determining unit calculates the distribution bias as:
The image processing apparatus according to claim 6, wherein the calculation is performed based on a difference between a luminance value of each pixel in the specific luminance region and an average luminance value in the specific luminance region in the luminance frequency distribution. 9. The method according to claim 8, wherein the threshold value determining unit calculates the distribution bias as:
9. The image processing apparatus according to claim 8, wherein the calculation is performed based on an odd power of a difference between a luminance value of each pixel in the specific luminance region and an average luminance value in the specific luminance region in the luminance frequency distribution. . 10. The threshold value determining means determines that the predetermined condition is satisfied if the distribution bias is within a predetermined range, and if the distribution bias is outside the predetermined range and is positive, the first Changing the luminance value to an average luminance value in the specific luminance region, and if the distribution bias is outside the predetermined range and negative, changing the second luminance value to an average luminance value in the specific luminance region. 7. The image processing apparatus according to claim 6, wherein: