JP7452120B2 - Image processing method, image processing program, and image processing device - Google Patents

Description

The present invention relates to an image processing method and the like.

A contract contains information about the person who has concluded the contract, such as the address and name, and is used for various processing. In the following explanation, information regarding the person who concluded the contract will be referred to as "contractor information" as appropriate.

Here, when detecting the conclusion party information written in a paper contract, a scanner is used to capture the contract as image data, and image processing for character detection is performed on the image data. Note that if image processing for character detection is performed on the entire image data of the contract, the image processing will also be performed on areas that do not include signatory information, reducing detection efficiency.

In addition, if the format of the contract is determined in advance, image processing for character detection can be performed on a predetermined area of the image data of the contract, improving the efficiency of detecting information on the signatory. Can be done.

Japanese Patent Application Publication No. 2001-184481

However, the above-described conventional technology has a problem in that it is not possible to efficiently extract the region in which the conclusion party information is written.

For example, if the format of the contract is unknown, it is not possible to identify in advance the area where the signatory information is written, and image processing for character detection is performed on the entire image data of the contract. As a result, detection efficiency decreases.

In one aspect, an object of the present invention is to provide an image processing method, an image processing program, and an image processing device that can efficiently extract a region in which conclusion person information is written.

In the first plan, the computer executes the following process. The computer processes image data generated by electronically reading a sealed contract document, and identifies an area corresponding to the stamp surface from the image data. The computer stores a first area in the contract corresponding to the stamp position of the contract and a second area in the contract in which information about the person who concluded the contract is written, which are stored in advance in the storage device. With reference to the positional relationship, a region specified based on the positional relationship with the region of the stamp surface specified from the image data to be processed is extracted from the image data to be processed. The computer outputs the extracted area.

It is possible to efficiently extract the area in which the conclusion party information is written.

FIG. 1 is a diagram for explaining the processing of the image processing apparatus according to the present embodiment. FIG. 2 is a functional block diagram showing the configuration of the image processing apparatus according to this embodiment. FIG. 3 is a diagram showing an example of the data structure of learning data. FIG. 4 is a diagram showing an example of the data structure of pattern data. FIG. 5 is a diagram showing a histogram of the first y-coordinate difference. FIG. 6 is a diagram showing an example of detection of the first pattern. FIG. 7 is a diagram showing an example of detection of the second pattern. FIG. 8 is a flowchart showing the processing procedure of the learning process of the image processing apparatus according to the present embodiment. FIG. 9 is a flowchart showing the procedure of image processing by the image processing apparatus according to this embodiment. FIG. 10 is a diagram showing a comparison result between the first pattern and the second pattern. FIG. 11 is a diagram illustrating an example of the hardware configuration of a computer that implements the same functions as the image processing apparatus of the embodiment.

Embodiments of an image processing method, an image processing program, and an image processing apparatus disclosed in the present application will be described in detail below based on the drawings. Note that the present invention is not limited to this example.

FIG. 1 is a diagram for explaining the processing of the image processing apparatus according to the present embodiment. The image processing device learns pattern data 142 based on learning data 141. The pattern data 142 is information indicating a detection area based on the center coordinates of the stamp face 10. For example, the detection area corresponds to an area of the contract where information about the person who concluded the contract (contractor information) is written. The center coordinates of the stamp face 10 (the region of the stamp face including the center coordinates) correspond to the “first region”.

The pattern data 142 defines a first y-coordinate difference (y ₁ ) and a second y-coordinate difference (y ₂ ) based on the center coordinates of the stamp face 10. The first y-coordinate difference is the difference between the upper end 11a of the detection area and the center coordinate of the stamp surface 10. The second y-coordinate difference is the difference between the lower end 11b of the detection area and the center coordinate of the stamp face 10. The area corresponding to the y coordinates of the upper end 11a and the lower end 11b corresponds to the "second area".

When the image processing device uses the pattern data 142 to extract the detection area 25 included in the image data 20 of the contract, the stamp face 20a of the image data 20 is specified. The image processing device specifies the upper end 25a of the detection area 25 based on the center coordinates of the stamp surface 20a and the first y-coordinate difference (y ₁ ). The image processing device specifies the lower end 25b of the detection area 25 based on the center coordinates of the stamp surface 20a and the second y-coordinate difference (y ₂ ). Note that the image processing device makes the width x25 of the detection area 25 in the x-coordinate direction the same as the width x20 of the image data 20. The image processing device recognizes the character string of the conclusion party information by performing character recognition on the detection area 25.

As described above, the image processing device according to the present embodiment identifies the stamp face 20a from the image data 20 of the contract, and based on the positional relationship of the detection area with respect to the central coordinates of the stamp face defined in the pattern data 142. Then, the detection area 25 of the image data 20 is extracted. In other words, even if the format of the contract is unknown, the image processing device can efficiently detect the detection area where the signatory information is written by focusing on the fact that the positional relationship between the seal position and the position where the signatory information is written does not change. can be extracted to Further, by efficiently extracting the detection area, it is also possible to appropriately recognize the conclusion party information that may be written in the detection area.

Next, the configuration of the image processing apparatus according to this embodiment will be explained. FIG. 2 is a functional block diagram showing the configuration of the image processing apparatus according to this embodiment. As shown in FIG. 2, this image processing device 100 is connected to a scanner 50. The image processing device 100 includes a communication section 110, an input section 120, a display section 130, a storage section 140, and a control section 150.

The scanner 50 is a device that generates image data of a contract by electronically reading the front surface of the contract. In this embodiment, it is assumed that the scanner 50 generates image data of a sealed contract. The scanner 50 outputs image data to the communication unit 110 of the image processing device 100.

The communication unit 110 is a device that connects to the scanner 50 and acquires image data. Communication unit 110 is an example of a communication device. Although not shown, the communication unit 110 may connect to another external device via a network and perform data communication. A control unit 150, which will be described later, exchanges data with the scanner 50, external devices, etc. via the communication unit 110.

The input unit 120 is an input device for inputting various data to the control unit 150 of the image processing apparatus 100. The input unit 120 corresponds to a keyboard, a mouse, a touch panel, etc.

The display unit 130 is a display device that displays information output from the control unit 150. The display unit 130 corresponds to an organic EL (Electro Luminescence) display, a liquid crystal display, a touch panel, or the like.

The storage unit 140 has learning data 141 and pattern data 142. The storage unit 140 corresponds to a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory, or a storage device such as an HDD (Hard Disk Drive).

Learning data 141 is data used to generate pattern data 142. FIG. 3 is a diagram showing an example of the data structure of learning data. As shown in FIG. 3, this learning data 141 includes a sample number, image data, upper end y-coordinate, lower end y-coordinate, stamp face center x-coordinate, stamp face center y-coordinate, first y-coordinate difference, and first y-coordinate. 2y coordinate difference.

The sample number is information that identifies image data serving as a sample. The image data is image data of a sample contract that has been digitized using a scanner or the like. The sample contract contains the information of the signatory and is stamped. The image data of the sample contract corresponds to "second image data." In the following description, the image data of the learning data will be referred to as second image data, especially when distinguished.

The upper end y-coordinate indicates the y-coordinate of the upper end of the detection area in the second image data. The lower end y-coordinate indicates the y-coordinate of the lower end of the detection area in the second image data. The seal surface center x-coordinate is the center x-coordinate of the stamp surface in the second image data. The stamp surface center y-coordinate is the center y-coordinate of the stamp surface in the second image data. For example, the coordinates of the upper left corner of the second image data are set as the origin.

The first y-coordinate difference is a difference value obtained by subtracting the top y-coordinate from the stamp face center y-coordinate. The second y-coordinate is a difference value obtained by subtracting the stamp face center coordinate from the lower end y-coordinate.

Returning to the explanation of FIG. 2. The pattern data 142 is information generated based on the learning data 141, and is information indicating a detection area based on the central coordinates of the stamp face. The detected coordinates of the pattern data 142 indicate an area in the image data of the new contract where the signatory information is likely to be written. FIG. 4 is a diagram showing an example of the data structure of pattern data. As shown in FIG. 4, the pattern data 142 includes a first pattern and a second pattern.

In the first pattern, a 1st-1y coordinate difference and a 2nd-1y coordinate difference are set. The 1-1st y-coordinate difference is the largest first y-coordinate difference among the first y-coordinate differences stored in the learning data 141. The 2-1st y-coordinate difference is the largest second y-coordinate difference among the second y-coordinate differences stored in the learning data 141.

In the second pattern, a 1st-2nd y coordinate difference and a 2nd-2nd y coordinate difference are set. The 1st-2nd y-coordinate difference is the largest first y-coordinate difference among the first y-coordinate differences other than the outliers, excluding outliers from each 1y-coordinate difference stored in the learning data 141. The 2-2nd y-coordinate difference is the largest second y-coordinate difference among the second y-coordinate differences other than the outliers, excluding outliers from each second y-coordinate difference stored in the learning data 141.

The second pattern excludes outliers. Therefore, the detection area specified by the first pattern is wider than the detection area specified by the second pattern.

Returning to the explanation of FIG. 2. The control unit 150 includes an acquisition unit 151 , a learning unit 152 , a specification unit 153 , an extraction unit 154 , and a character recognition unit 155 . The control unit 150 can be realized by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like. Further, the control unit 150 can also be realized by hardwired logic such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

The acquisition unit 151 is a processing unit that acquires the learning data 141 from an external device or the like. The acquisition unit 151 stores the acquired learning data 141 in the storage unit 140. Further, when the acquisition unit 151 acquires image data of a new contract with a stamp from the scanner 50, the acquisition unit 151 outputs the acquired image data to the identification unit 153. The image data of the new signed contract corresponds to the "first image data." In the following description, the image data of a new contract that has been stamped will be referred to as first image data, especially if it is to be distinguished.

The learning unit 152 is a processing unit that generates pattern data 142 based on the learning data 141. Note that the seal surface center y-coordinate (seal surface center x-coordinate) included in the learning data 141 may be set in advance, or may be automatically set by specifying the stamp surface from the second image data. When the learning unit 152 automatically specifies the stamp face center y-coordinate, it may automatically calculate the first y-coordinate difference and the second y-coordinate difference based on the upper end y-coordinate and the lower end y-coordinate.

A process in which the learning unit 152 learns the "first pattern" of the pattern data 142 will be described. The learning unit 152 compares each first y-coordinate difference of the learning data 141 and identifies the largest first y-coordinate difference. The learning unit 152 learns the identified maximum first y coordinate difference as the 1-1 y coordinate difference of the first pattern.

The learning unit 152 compares each second y-coordinate difference of the learning data 141 and identifies the maximum second y-coordinate difference. The learning unit 152 learns the identified maximum second y-coordinate difference as the 2-1 y-coordinate difference of the first pattern.

The learning unit 152 registers the 1st-1y coordinate difference and the 2nd-1y coordinate difference in the pattern data 142 as a first pattern.

Next, a process in which the learning unit 152 learns the "second pattern" of the pattern data 142 will be described. The learning unit 152 calculates the average value μ ₁ of the first y-coordinate differences based on all the first y-coordinate differences of the learning data 141. Furthermore, the learning unit 152 calculates the standard deviation σ ₁ of the first y-coordinate differences based on all the first y-coordinate differences of the learning data 141.

Of all the first y-coordinate differences of the learning data 141, the learning unit 152 selects a first y-coordinate difference that is less than “μ ₁ −3×σ ₁ ” and a first y-coordinate difference that is greater than or equal to “μ ₁ +3×σ ₁ ”. The difference is identified as an outlier. The learning unit 152 learns the largest first y-coordinate difference among all the first y-coordinate differences of the learning data 141 excluding outliers as the 2-1 y-coordinate difference of the second pattern.

FIG. 5 is a diagram showing a histogram of the first y-coordinate difference. The horizontal axis of the histogram shown in FIG. 5 corresponds to the value of the first y-coordinate difference. The vertical axis of the histogram corresponds to frequency. For example, the value of the first y-coordinate difference whose value is included in the range 30 becomes an outlier. For example, the value shown at 31a is the value of the 1-1y coordinate difference of the first pattern. The value shown in 31b is the value of the 1st-2nd y coordinate difference of the second pattern.

The learning unit 152 calculates the average value μ ₂ of the second y-coordinate differences based on all the second y-coordinate differences of the learning data 141. Further, the learning unit 152 calculates the standard deviation σ ₂ of the first y-coordinate difference based on all the second y-coordinate differences of the learning data 141.

The learning unit 152 selects a second y-coordinate difference that is less than “μ ₂ −3×σ ₂ ” and a second y-coordinate difference that is equal to or more than “μ ₂ +3×σ ₂ ” out of all the second y-coordinate differences of the learning data 141. The difference is identified as an outlier. The learning unit 152 learns the largest second y-coordinate difference among all the second y-coordinate differences of the learning data 141 excluding outliers as the 2-2nd y-coordinate difference of the second pattern. Regarding the first y-coordinate difference, explanation of the histogram will be omitted.

The learning unit 152 registers the 1st-2nd y coordinate difference and the 2nd-2nd y coordinate difference in the pattern data 142 as a second pattern.

The learning unit 152 learns the pattern data 142 by performing the above processing, and stores the learned pattern data 142 in the storage unit 140.

Returning to the explanation of FIG. 2. The specifying unit 153 is a processing unit that specifies the area coordinates corresponding to the stamp surface from the first image data. For example, the specifying unit 153 specifies the area of the stamp face by comparing the first image data with a template that defines the color and shape of the stamp. The specifying unit 153 specifies the center coordinates (x ₀ , y ₀ ) of the specified region of the seal face, and outputs information on the center coordinates (x ₀ , y ₀ ) to the extracting unit 154 . However, the area coordinates are not necessarily limited to the center coordinates; for example, they may be coordinates corresponding to the vertices of the stamp face, or a plurality of coordinates may be used as the area coordinates.

The extraction unit 154 is a processing unit that uses the pattern data 142 to extract a detection area included in the first image data. The extraction unit 154 extracts a detection area using either the first pattern or the second pattern included in the pattern data 142. It is assumed that whether to use the first pattern or the second pattern is set in advance.

An example of a process in which the extraction unit 154 extracts a detection area using the first pattern will be described. The extraction unit 154 specifies the upper end of the detection area 25 based on the central coordinates (x ₀ , y ₀ ) of the stamp face and the 1-1y coordinate difference. The extraction unit 154 specifies the lower end of the detection area based on the central coordinates (x ₀ , y ₀ ) of the stamp face and the 2-1y coordinate difference. Note that the extraction unit 154 sets the width of the detection area in the x-coordinate direction to be the same as the width of the first image data.

The extraction unit 154 outputs information on the detection area specified for the first image data to the character recognition unit 155 using the first pattern.

FIG. 6 is a diagram showing an example of detection of the first pattern. As shown in FIG. 6, a detection area 40a is set based on the center coordinates of the stamp surface 40 and the first pattern.

An example of a process in which the extraction unit 154 extracts a detection area using the second pattern will be described. The extraction unit 154 specifies the upper end of the detection area 25 based on the center coordinates (x ₀ , y ₀ ) of the stamp face and the 1st-2nd y coordinate difference. The extraction unit 154 specifies the lower end of the detection area based on the central coordinates (x ₀ , y ₀ ) of the stamp face and the 2-2nd y coordinate difference. Note that the extraction unit 154 sets the width of the detection area in the x-coordinate direction to be the same as the width of the first image data.

The extraction unit 154 outputs information on the detection area specified for the first image data to the character recognition unit 155 using the second pattern.

FIG. 7 is a diagram showing an example of detection of the second pattern. As shown in FIG. 7, a detection area 40b is set based on the center coordinates of the stamp surface 40 and the second pattern.

The character recognition unit 155 is a processing unit that recognizes the character string of the conclusion party information by performing character recognition on the image of the detection area extracted by the extraction unit 154. The character recognition unit 155 outputs the recognized character string to the display unit 130. The character recognition unit 155 may notify an external device or the like of the recognition result of the character string via the network.

Next, an example of the processing procedure of the image processing apparatus 100 according to the present embodiment will be described. FIG. 8 is a flowchart showing the processing procedure of the learning process of the image processing apparatus according to the present embodiment. As shown in FIG. 8, the acquisition unit 151 of the image processing device 100 acquires learning data 141 (step S101).

The learning unit 152 of the image processing device 100 compares each first y-coordinate difference of the learning data 141, and sets the largest first y-coordinate difference as the 1-1st coordinate difference (step S102). The learning unit 152 compares each second y-coordinate difference of the learning data 141, and sets the largest second y-coordinate difference as the 2-1st coordinate difference (step S103).

The learning unit 152 registers the first-1y coordinate difference and the second-1y coordinate difference as the first pattern in the pattern data 142 (step S104).

The learning unit 152 calculates the average value μ ₁ and the standard deviation σ ₁ based on each first coordinate difference of the learning data 141 (step S105). The learning unit 152 excludes outliers from each first y-coordinate difference of the learning data 141 based on the average value μ ₁ and the standard deviation σ ₁ (step S106).

The learning unit 152 sets the largest first y-coordinate difference among the first y-coordinate differences excluding outliers as the 1-2nd y-coordinate difference (step S107).

The learning unit 152 calculates the average value μ ₂ and the standard deviation σ ₂ based on each second coordinate difference of the learning data 141 (step S108). The learning unit 152 excludes outliers from each second y-coordinate difference of the learning data 141 based on the average value μ ₂ and standard deviation σ ₂ (step S109).

The learning unit 152 sets the largest second y-coordinate difference among the second y-coordinate differences excluding outliers as the 2-2nd y-coordinate difference (step S110).

The learning unit 152 registers the 1st-2nd y coordinate difference and the 2nd-2nd y coordinate difference in the pattern data 142 as a second pattern (step S111).

FIG. 9 is a flowchart showing the procedure of image processing by the image processing apparatus according to this embodiment. As shown in FIG. 9, the acquisition unit 151 of the image processing device 100 acquires first image data from the scanner 50 (step S201).

The specifying unit 153 of the image processing device 100 specifies the region of the seal face from the first image data (step S202). The specifying unit 153 specifies the center coordinates of the seal face (step S203).

If the extraction unit 154 of the image processing device 100 selects the first pattern (step S204, Yes), the process proceeds to step S205. On the other hand, when the extraction unit 154 does not select the first pattern (step S204, No), the extraction unit 154 moves to step S207.

Step S205 will be explained. The extraction unit 154 acquires the 1st-1y coordinate difference and the 2nd-1y coordinate difference of the first pattern from the pattern data 142 (step S205). The extraction unit 154 extracts a detection area based on the first pattern and the center coordinates of the seal face (step S206).

The character recognition unit 155 of the image processing device 100 recognizes a character string from the image of the detection area (step S209). The character recognition unit 155 outputs the recognition result to the display unit 130 (step S210).

Step S207 will be explained. The extraction unit 154 acquires the 1st-2nd y coordinate difference and the 2nd-2nd y coordinate difference of the second pattern from the pattern data 142 (step S207). The extraction unit 154 extracts a detection area based on the second pattern and the center coordinates of the seal face (step S208), and proceeds to step S209.

Next, the effects of the image processing apparatus 100 according to the first embodiment will be explained. The image processing device 100 identifies the seal face from the image data of the contract, and extracts the detection area of the image data based on the positional relationship of the detection area with respect to the central coordinates of the seal face defined in the pattern data. In other words, even if the format of the contract is unknown, the image processing device can efficiently detect the detection area where the signatory information is written by focusing on the fact that the positional relationship between the seal position and the position where the signatory information is written does not change. can be extracted to Further, by efficiently extracting the detection area, it is also possible to appropriately recognize the conclusion party information that may be written in the detection area.

The image processing device 100 learns the first pattern based on the maximum value of each first y-coordinate difference and the maximum value of the second y-coordinate difference included in the learning data 141. With this, by extracting the seal face from the image data, it becomes possible to specify the detection area of the signatory information.

The image processing device 100 calculates the remaining first y-coordinate differences after excluding outliers from each first y-coordinate difference included in the learning data 141 and the remaining second y-coordinate differences after excluding outliers from each second y-coordinate difference. Based on this, a second pattern is learned. With this, by extracting the seal face from the image data, it becomes possible to specify the detection area of the signatory information. Furthermore, the detection area can be narrowed down more than the detection area of the first pattern.

Here, an example of recognition results when character recognition is performed on the detection area specified by the first pattern and the detection area specified by the second pattern will be described. FIG. 10 is a diagram showing a comparison result between the first pattern and the second pattern.

In the example shown in FIG. 10, the detection area 45a is set based on the center coordinates of the stamp surface 45 and the first pattern. A detection area 46a is set based on the center coordinates of the stamp surface 46 and the first pattern. In the example shown in FIG. 10, in the first pattern, the conclusion party information is detected, but in the second pattern, part of the conclusion party information cannot be detected (some detection is omitted). Since the detection area of the first pattern is larger than the detection area of the second pattern, detection omissions can be prevented.

For example, when the inventor conducted an experiment, the detection area of the first pattern was approximately 40% of the form, and the detection rate of the conclusion party information was 100%. On the other hand, the detection area of the second pattern was approximately 1.50% of the form, and the detection rate of the conclusion party information was 97.8%.

Next, an example of the hardware configuration of a computer that implements the same functions as the image processing apparatus 100 shown in the above embodiment will be described. FIG. 11 is a diagram illustrating an example of the hardware configuration of a computer that implements the same functions as the image processing apparatus of the embodiment.

As shown in FIG. 11, the computer 200 includes a CPU 201 that executes various calculation processes, an input device 202 that receives data input from a user, and a display 203. The computer 200 also includes a reading device 204 that reads programs and the like from a storage medium, and a communication device 205 that exchanges data with the scanner 50, external devices, etc. via a wired or wireless network. The computer 200 also includes a RAM 206 that temporarily stores various information and a hard disk device 207. Each device 201-207 is then connected to a bus 208.

The hard disk device 207 includes an acquisition program 207a, a learning program 207b, a specific program 207c, an extraction program 207d, and a character recognition program 207e. Further, the CPU 201 reads each program 207a to 207e and expands it into the RAM 206.

The acquisition program 207a functions as an acquisition process 206a. The learning program 207b functions as a learning process 206b. The specific program 207c functions as a specific process 206c. The extraction program 207d functions as an extraction process 206d. Character recognition program 207e functions as character recognition process 206e.

The processing of the acquisition process 206a corresponds to the processing of the acquisition unit 151. The processing of the learning process 206b corresponds to the processing of the learning unit 152. The processing of the identification process 206c corresponds to the processing of the identification unit 153. The processing of the extraction process 206d corresponds to the processing of the extraction unit 154. The processing of the character recognition process 206e corresponds to the processing of the character recognition unit 155.

Note that each of the programs 207a to 207e does not necessarily have to be stored in the hard disk device 207 from the beginning. For example, each program is stored in a "portable physical medium" such as a flexible disk (FD), CD-ROM, DVD, magneto-optical disk, or IC card that is inserted into the computer 200. Then, the computer 200 may read and execute each program 207a to 207e.

Regarding the embodiments including each of the above examples, the following additional notes are further disclosed.

(Additional note 1) On the computer,
Targeting image data generated by electronically reading a sealed contract document, identifying an area corresponding to the stamp surface from the image data,
The positional relationship between a first area in the contract corresponding to the stamp position of the contract and a second area in the contract in which information regarding the person who concluded the contract is written, which is stored in advance in the storage device. extracting a region specified based on the region of the seal face and the positional relationship identified from the image data to be processed by referring to the image data to be processed;
An image processing method characterized by executing a process of outputting the extracted region.

(Additional Note 2) The positions of the first area and the second area are determined based on learning data having multiple image data generated by electronically reading a sealed contract document. The image processing method according to appendix 1, further comprising the step of learning the relationship and storing the learning result in the storage device.

(Additional note 3) The learning process calculates a difference value between the coordinates specified as the first area and the coordinates specified as the second area for each image data, and among the calculated difference values. , the maximum difference value is learned as the positional relationship, the image processing method according to appendix 2.

(Additional note 4) The learning process calculates a difference value between the coordinates specified as the first area and the coordinates specified as the second area for each image data, and calculates a plurality of calculated difference values. The image according to appendix 2, wherein an outlier of the difference values is calculated based on the difference value, and the positional relationship is learned based on the difference value obtained by excluding the outlier from the plurality of difference values. Processing method.

(Additional note 5) On the computer,
Targeting image data generated by electronically reading a sealed contract document, identifying an area corresponding to the stamp surface from the image data,
The positional relationship between a first area in the contract corresponding to the stamp position of the contract and a second area in the contract in which information regarding the person who concluded the contract is written, which is stored in advance in the storage device. extracting a region specified based on the region of the seal face and the positional relationship identified from the image data to be processed by referring to the image data to be processed;
An image processing program that causes a process to output the extracted area.

(Additional Note 6) The positions of the first area and the second area are determined based on learning data having multiple image data generated by electronically reading a sealed contract document. The image processing program according to appendix 5, further comprising the step of learning a relationship and storing the learning result in the storage device.

(Additional Note 7) The learning process calculates a difference value between the coordinates specified as the first area and the coordinates specified as the second area for each image data, and among the calculated difference values. , the maximum difference value is learned as the positional relationship, the image processing program according to appendix 6.

(Additional note 8) The learning process calculates a difference value between the coordinates specified as the first area and the coordinates specified as the second area for each image data, and calculates a plurality of calculated difference values. The image according to appendix 6, wherein an outlier of the difference values is calculated based on the difference value, and the positional relationship is learned based on the difference value obtained by excluding the outlier from the plurality of difference values. Processing program.

(Additional note 9) A specifying unit that specifies an area corresponding to the stamp surface from the image data, with image data generated by electronically reading a sealed contract document as a processing target;
The positional relationship between a first area in the contract corresponding to the stamp position of the contract and a second area in the contract in which information regarding the person who concluded the contract is written, which is stored in advance in the storage device. an extraction unit that refers to and extracts a region specified from the image data that is a processing target based on the area of the stamp face specified from the image data that is a processing target and the positional relationship, and outputs the extracted region; An image processing device comprising: and

(Additional Note 10) The positions of the first area and the second area are determined based on learning data having a plurality of image data generated by electronically reading a sealed contract document. The image processing device according to appendix 9, further comprising a learning section that learns the relationship and stores the learning result in the storage device.

(Additional Note 11) The learning unit calculates a difference value between the coordinates specified as the first region and the coordinates specified as the second region for each image data, and among the calculated difference values, The image processing device according to appendix 10, wherein the maximum difference value is learned as the positional relationship.

(Additional Note 12) The learning unit calculates a difference value between the coordinates specified as the first region and the coordinates specified as the second region for each image data, and calculates a plurality of calculated difference values. 11. The image according to claim 10, wherein an outlier of difference values is calculated based on the difference value, and the positional relationship is learned based on the difference value obtained by excluding the outlier from the plurality of difference values. Processing equipment.

50 Scanner 100 Image processing device 110 Communication unit 120 Input unit 130 Display unit 140 Storage unit 141 Learning data 142 Pattern data 150 Control unit 151 Acquisition unit 152 Learning unit 153 Specification unit 154 Extraction unit 155 Character recognition unit

Claims (6)

An image processing method performed by a computer, the method comprising:
Targeting image data generated by electronically reading a sealed contract document, identifying an area corresponding to the stamp surface from the image data,
A first area in the contract corresponding to the stamp position of the contract and a second area in the contract in which information regarding the person who concluded the contract is written, which are stored in advance in the storage device, The positional relationship between the area of the stamp surface and the area identified from the image data to be processed by referring to the positional relationship with the second area defined by two coordinate differences in the vertical direction with respect to the first area. extracting a region identified based on the above from the image data to be processed;
An image processing method characterized by executing a process of outputting the extracted region. The positional relationship between the first area and the second area is learned based on learning data having a plurality of image data generated by electronically reading a sealed contract document. 2. The image processing method according to claim 1, further comprising: storing the learning results in the storage device. The learning process calculates the difference value between the coordinates specified as the first area and the coordinates specified as the second area for each image data, and calculates the maximum difference among the calculated difference values. 3. The image processing method according to claim 2, wherein a value is learned as the positional relationship. The learning process calculates a difference value between the coordinates specified as the first region and the coordinates specified as the second region for each image data, and based on the plurality of calculated difference values. 3. The image processing method according to claim 2, further comprising calculating an outlier among the difference values, and learning the positional relationship based on the difference value obtained by excluding the outlier from the plurality of difference values. Processing target is image data generated by electronically reading a sealed contract document into a computer, and identifying an area corresponding to the stamp surface from the image data,
a first area in the contract corresponding to the stamp position of the contract and a second area in the contract in which information regarding the person who concluded the contract is written, which are stored in advance in the storage device; The positional relationship between the area of the stamp surface and the area identified from the image data to be processed by referring to the positional relationship with the second area defined by two coordinate differences in the vertical direction with respect to the first area. extracting a region identified based on from the image data to be processed;
An image processing program that causes a process to output the extracted area. a specifying unit for processing image data generated by electronically reading a sealed contract document, and identifying an area corresponding to the stamp surface from the image data;
a first area in the contract corresponding to the stamp position of the contract and a second area in the contract in which information regarding the person who concluded the contract is written, which are stored in advance in the storage device; The positional relationship between the area of the stamp surface and the area identified from the image data to be processed by referring to the positional relationship with the second area defined by two coordinate differences in the vertical direction with respect to the first area. an extraction unit that extracts a region specified based on the image data to be processed from the image data to be processed, and outputs the extracted region.

Images