阅读说明：本技术 处理装置、处理方法和程序 (Processing device, processing method, and program ) 是由 马场英一 千叶博司 于 2020-04-24 设计创作，主要内容包括：本发明提供了一种处理装置(10),包括：获取单元(11),用于获取填入表单的图像,填入表单具有填入数值的多个第一填入字段和填入在多个第一填入字段中被填入的数值的合计的第二填入字段；分析单元(12),用于分析所述图像,并识别已被填入在多个第一填入字段中的值和已被填入在第二填入字段中的值；确定单元(13),确定已被填入在多个第一填入字段中的值的识别结果的合计与已被填入在第二填入字段中的值的识别结果是否彼此匹配；以及处理单元(14),如果已被填入在多个第一填入字段中的值的识别结果的合计与已被填入在第二填入字段中的值的识别结果彼此不匹配时,执行错误处理。(The invention provides a processing apparatus (10) comprising: an acquisition unit (11) for acquiring an image of a fill-in form having a plurality of first fill-in fields filled with numerical values and a second fill-in field filled with a total of the numerical values filled in the plurality of first fill-in fields; an analyzing unit (12) for analyzing the image and identifying values that have been filled in a plurality of first filled-in fields and values that have been filled in second filled-in fields; a determination unit (13) that determines whether the sum of the recognition results of the values that have been filled in the plurality of first fill-in fields and the recognition result of the value that has been filled in the second fill-in field match each other; and a processing unit (14) that performs error processing if the sum of the identification results of the values that have been filled in the plurality of first filled-in fields and the identification result of the value that has been filled in the second filled-in field do not match each other.)

1. A processing apparatus, comprising:

an acquisition device configured to acquire an image of a fill-in form, the fill-in form including a plurality of first fill-in fields filled with numerical values and a second fill-in field filled with a total of the numerical values filled in the plurality of first fill-in fields;

analyzing means for analyzing the image and identifying a value filled in a plurality of the first filled fields and a value filled in the second filled field;

a determination means for determining whether or not an aggregate of the recognition results of the values filled in the plurality of first filled fields and the recognition result of the value filled in the second filled field match each other; and

processing means for performing error processing when the sum of the identification results of the values filled in the plurality of first filled fields and the identification result of the value filled in the second filled field do not match each other.

2. The processing apparatus according to claim 1,

the fill-in form is a transfer request form, an

The first fill-in field is a fill-in field to fill in the transfer amount.

3. The processing apparatus according to claim 1 or 2,

the fill-in form includes a plurality of third fill-in fields that fill in the same value,

the analyzing means analyzes the image and identifies a value filled in each of a plurality of the third filled fields,

the determination means determines whether the recognition results of the values filled in the plurality of third fill-in fields match each other, and

the processing means performs error processing when the identification results of the values filled in the plurality of third filled fields do not match each other.

4. The processing apparatus according to claim 3,

the fill-in form is a transfer request form, an

A plurality of the third fill-in fields includes a fill-in field for a transfer requester fill-in value and a fill-in field for a staff fill-in value.

5. The processing apparatus according to any one of claims 1 to 4,

the fill-in form is a transfer request form and includes an item fill-in field for filling in items and a specified transfer date fill-in field for filling in a specified transfer date,

the analyzing means analyzes the image and identifies a value filled in the item fill-in field and a value filled in the specified transfer date fill-in field,

the determination means determines whether the result of identification of the value filled in the entry fill-in field and the result of identification of the value filled in the specified transfer date fill-in field satisfy the condition based on condition information that determines the condition of the specifiable specified transfer date for each of the entries, and

the processing means performs error processing when the recognition result of the value filled in the entry fill-in field and the recognition result of the value filled in the specified transfer date fill-in field do not satisfy the condition.

6. The processing apparatus according to any one of claims 1 to 5,

the fill-in form includes a date fill-in field to fill in the date and a day fill-in field to fill in the day,

the analyzing means analyzes the image and identifies a value filled in the date filled-in field and a value filled in the day filled-in field,

the determination means determines whether or not the identification result of the value filled in the date filling field and the identification result of the value filled in the day filling field are contradictory based on the calendar information indicating the date and the day of week, and

the processing means performs error processing when the identification result of the value filled in the date filling field contradicts the identification result of the value filled in the day filling field.

7. The processing apparatus according to claim 6,

the fill-in form is a transfer request form, an

The date fill-in field is a fill-in field that fills in the specified transfer date.

8. The processing apparatus according to any one of claims 1 to 7,

the analysis means analyzes the image using a plurality of image analysis engines and identifies values filled in a plurality of fill-in fields included in the fill-in form, and

when the recognition results of the plurality of image analysis engines do not match each other, the processing means performs error processing.

9. A processing apparatus, comprising:

an acquisition means for acquiring an image of a transfer request form including a plurality of fill-in fields;

analyzing means for identifying values in the plurality of fill-in fields included in the transfer request form using a plurality of image analysis engines;

search means for searching the storage unit using, as a keyword, a recognition result of a value filled in a keyword filling field of each of the plurality of image analysis engines;

determining means for determining whether there is a value satisfying an error condition in recognition results of values in a plurality of fill-in fields by a plurality of the image analysis engines; and

processing means for performing error processing when recognition results of a plurality of the image analysis engines do not match each other, when the keyword generated by any one of a plurality of the image analysis engines cannot be found in the storage unit, or when a value satisfying the error condition exists in recognition results of a plurality of the image analysis engines on values in a plurality of filled-in fields.

10. A method of processing, comprising:

by means of the computer, it is possible to,

acquiring an image of a filling form, wherein the filling form comprises a plurality of first filling fields filled with numerical values and a second filling field filled with the total numerical values filled in the first filling fields;

analyzing the image and identifying values filled in a plurality of the first filled fields and values filled in the second filled fields;

determining whether an aggregate of the recognition results of the values filled in the plurality of first filled fields and the recognition result of the value filled in the second filled field match each other; and

performing error processing when the total of the recognition results of the values filled in the plurality of first filled fields and the recognition result of the value filled in the second filled field do not match each other.

11. A program that causes a computer to function as:

an acquisition device configured to acquire an image of a fill-in form, the fill-in form including a plurality of first fill-in fields filled with numerical values and a second fill-in field filled with a total of the numerical values filled in the plurality of first fill-in fields;

analyzing means for analyzing the image and identifying a value filled in a plurality of the first filled fields and a value filled in the second filled field;

a determination means for determining whether or not an aggregate of the recognition results of the values filled in the plurality of first filled fields and the recognition result of the value filled in the second filled field match each other; and

processing means for performing error processing when the sum of the identification results of the values filled in the plurality of first filled fields and the identification result of the value filled in the second filled field do not match each other.

Technical Field

The invention relates to a processing apparatus, a processing method, and a program.

Background

Patent document 1 discloses a system that leaves a display area of characters on a correction screen blank and causes a confirmer to input characters when a customer fill-in amount and a recognition result of an operator's repeatedly-credited amount do not match.

Patent document 2 discloses a financial system formed by connecting a terminal device operated by an operator to a server that performs data processing through a network. In the financial system, for each of a plurality of documents, an aggregation document including at least data indicating an aggregation of amounts of money processed in each document and a priority degree of processing is generated, and the data of the aggregation document is stored in a server. Then, the server receives at least an input of data of the amount of money from each document compiled by the aggregation document, calculates the aggregation, and determines the amount of money by checking the stored data of the aggregation document. Further, the server determines the priority of processing from the data of the aggregated documents, and allocates the processing to the terminal device.

[ relevant documents ]

[ patent document ]

[ patent document 1] Japanese patent application laid-open No.2001-76093

[ patent document 2] Japanese patent application laid-open No. H11-328285

Disclosure of Invention

[ problem ] to

It is practiced to identify the content filled in the transfer form by using Optical Character Recognition (OCR). However, the recognition accuracy of the OCR is not 100%, there is a possibility of erroneous recognition, and therefore, a manual confirmation work is required before the transfer process is performed. It is desirable to reduce the burden of manual validation work. The present invention challenges to provide techniques for detecting misrecognitions by OCR.

[ technical solution ]

The present invention provides a processing apparatus, including:

an acquisition means for acquiring an image of a fill-in form including a plurality of first fill-in fields to which numerical values are filled and a second fill-in field to which a total of the numerical values filled in the plurality of first fill-in fields is filled;

analyzing means for analyzing the image and identifying a value filled in the plurality of first filled-in fields and a value filled in the second filled-in field;

determining means for determining whether or not the sum of the recognition results of the values filled in the plurality of first filled-in fields and the recognition result of the value filled in the second filled-in field match each other; and

processing means for performing error processing when the sum of the identification results of the values filled in the plurality of first filled-in fields and the identification result of the value filled in the second filled-in field do not match each other.

Further, the present invention provides a processing apparatus comprising:

an acquisition means for acquiring an image of a transfer request form including a plurality of fill-in fields;

analyzing means for identifying values in a plurality of filled-in fields included in the transfer request form using a plurality of image analysis engines;

search means for searching the storage unit using, as a keyword, a recognition result of a value filled in a keyword filling field of each of the plurality of image analysis engines;

determining means for determining whether there is a value satisfying an error condition in recognition results of the values in the plurality of filled-in fields by the plurality of image analysis engines; and

processing means for performing error processing when the recognition results of the plurality of image analysis engines do not match each other, when a keyword generated by each of the plurality of image analysis engines cannot be found in the storage unit, or when a value satisfying an error condition exists in the recognition results of the plurality of image analysis engines for values in the plurality of fill-in fields.

Furthermore, the present invention provides a processing method comprising:

by means of the computer, it is possible to,

acquiring an image of a filling form, wherein the filling form comprises a plurality of first filling fields filled with numerical values and a second filling field filled with the total numerical values filled in the plurality of first filling fields;

analyzing the image and identifying values filled in the plurality of first filled fields and values filled in the second filled fields;

determining whether an aggregate of the recognition results of the values filled in the plurality of first filled-in fields and the recognition result of the value filled in the second filled-in field match each other; and

when the sum of the recognition results of the values filled in the plurality of first filled-in fields and the recognition result of the value filled in the second filled-in field do not match each other, error processing is performed.

Further, the present invention provides a program that causes a computer to function as:

an acquisition device for acquiring an image of a fill-in form including a plurality of first fill-in fields to which numerical values are filled and a second fill-in field to which a total of the numerical values filled in the plurality of first fill-in fields is filled;

analyzing means for analyzing the image and identifying a value filled in the plurality of first filled fields and a value filled in the second filled field;

determining means for determining whether or not the sum of the recognition results of the values filled in the plurality of first filled-in fields and the recognition result of the value filled in the second filled-in field match each other; and

processing means for performing error processing when the sum of the identification results of the values filled in the plurality of first filled-in fields and the identification result of the value filled in the second filled-in field do not match each other.

[ Effect of the invention ]

The present invention implements a technique for detecting misrecognition by OCR.

Drawings

The above objects, other objects, features and advantages will become more apparent from the following description of preferred exemplary embodiments and the following drawings.

Fig. 1 is a diagram illustrating an overview of a processing apparatus according to the present exemplary embodiment.

Fig. 2 is a diagram illustrating an overview of a processing apparatus according to the present exemplary embodiment.

Fig. 3 is a diagram illustrating one example of a hardware configuration of a processing apparatus according to the present exemplary embodiment.

Fig. 4 is one example of a functional block diagram of a processing apparatus according to the present exemplary embodiment.

Fig. 5 is a flowchart illustrating one example of a processing flow in the processing apparatus according to the present exemplary embodiment.

Fig. 6 is a diagram illustrating one example of a fill-in form according to the present exemplary embodiment.

Fig. 7 is a diagram schematically illustrating one example of information processed by the processing apparatus according to the present exemplary embodiment.

Fig. 8 is one example of a functional block diagram of a processing apparatus according to the present exemplary embodiment.

Fig. 9 is a flowchart illustrating one example of a processing flow in the processing apparatus according to the present exemplary embodiment.

Fig. 10 is a diagram schematically illustrating one example of information processed by the processing apparatus according to the present exemplary embodiment.

Detailed Description

< first exemplary embodiment >

First, an outline of a processing apparatus according to the present exemplary embodiment is described. The processing apparatus according to the present exemplary embodiment is used in a financial institution such as a bank.

As shown in fig. 1, the processing device 10 cooperates with a settlement host 20 that performs transfer processing. First, before the transfer process of the settlement host 20 is performed, an image (image data) of a transfer request form (fill-in form) generated by an image scanner or the like is input to the processing device 10. An overview of the process flow of the processing apparatus 10 is illustrated in fig. 2.

First, the processing device 10 recognizes the contents filled in the transfer request form (image) read and input by the OCR, and converts the contents into digital character codes. Next, the processing device 10 determines whether the recognition result satisfies a predetermined error condition.

When the error condition is not satisfied, the processing device 10 inputs the recognition result (transfer request content) to the settlement host 20. The settlement host 20 performs transfer processing based on the inputted recognition result. On the other hand, when the error condition is satisfied, the processing device 10 executes error processing. Error handling is any process for performing a manual validation job. Thereafter, the recognition result (transfer request content) corrected by the manual confirmation work or the like is input to the settlement host 20. The settlement host 20 performs transfer processing based on the input contents. The processing apparatus 10 according to the present exemplary embodiment accurately detects the misrecognition of the OCR by determining whether a character error condition is satisfied. Details are described later.

Now, the configuration of the processing apparatus 10 is described in detail. First, one example of the hardware configuration of the processing apparatus 10 is described. Each function included in the processing device 10 is realized by any combination of hardware and software, which mainly includes a Central Processing Unit (CPU) of any computer, a memory, a program loaded onto the memory, a storage unit such as a hard disk storing the program (which can store not only the program stored in advance from the stage of the transportation device but also the program downloaded from a storage medium such as a Compact Disc (CD) or a server on the internet, etc.), and an interface for network connection. Those skilled in the art will appreciate, however, that there are numerous alternative examples of methods and apparatus for implementing.

Fig. 3 is a block diagram showing a hardware configuration of the processing apparatus 10. As shown in fig. 3, the processing device 10 includes a processor 1A, a memory 2A, an input/output interface 3A, a peripheral circuit 4A, and a bus 5A. The peripheral circuit 4A includes various modules. The processing apparatus 10 may not include the peripheral circuit 4A. Note that the processing device 10 may be configured by one physically and/or logically integrated device, or may be configured by a plurality of physically and/or logically separate devices. When the processing device 10 is configured by a plurality of physically and/or logically separated devices, each of the plurality of devices may include the above-described hardware configuration.

The bus 5A is a data transmission path for the processor 1A, the memory 2A, the peripheral circuit 4A, and the input/output interface 3A to transmit and receive data to and from each other. The processor 1A is, for example, an arithmetic processing device such as a CPU and a Graphics Processing Unit (GPU). The memory 2A is, for example, a memory such as a Random Access Memory (RAM) and a Read Only Memory (ROM). The input/output interface 3A includes an interface for acquiring information from an input device, an external server, an external sensor, a camera, and the like, an interface for outputting information to an output device, an external server, and the like. The input device is, for example, a keyboard, a mouse, a microphone, a touch panel, a physical button, a camera, or the like. Output devices are, for example, displays, speakers, printers, mailers, etc. The processor 1A may give an instruction to each module and perform an arithmetic operation based on an arithmetic result of each module.

Now, the functional configuration of the processing apparatus 10 is described in detail by using the functional block diagram in fig. 4 and the flowchart in fig. 5. One example of a functional block diagram of the processing device 10 is illustrated in fig. 4. As shown, the processing device 10 includes an acquisition unit 11, an analysis unit 12, a determination unit 13, and a processing unit 14.

First, in S10, the acquisition unit 11 acquires an image of a filled-in form generated by an image scanner or the like. In the present exemplary embodiment, the fill form is a transfer request form. An example of a transfer request form is shown in FIG. 6. Note that the illustrated transfer request form is only one example, and the present invention is not limited to the transfer request form.

The illustrated transfer request form includes the following fill fields, and the like.

"field filled with information identifying transfer requester (customer number and name of transfer requester)"

"fill in field of acceptance date"

"fill in field specifying transfer date"

"field to fill in the day of the week specifying the transfer date"

"filling in fields of items"

"fields to fill in document sequence number"

"fields filled with transfer destination information (financial institution name and branch name), payee information (item, account number and name), and transfer content (transfer amount) for each transfer request (eight cases)"

"fill in a subtotal field for transfer amount"

"field to fill in the totality of the sub-total of the multiple transfer request forms"

' Small ' record-copy field '

In this manner, the transfer request form includes a plurality of first fill fields (the "amount" field shown in fig. 6) to fill in the numerical value and a second fill field (the "subtotal field" shown in fig. 6) to fill in the aggregate of the numerical values filled in the plurality of first fill fields.

Note that in this specification, "acquiring" includes at least any one of the following: "data that is extracted by the local device (actively acquired) from data stored in another device or storage medium", for example, received by requesting or querying another device, accessed and read from another device or storage medium based on user input or instructions based on a program, or the like; "input data output from another apparatus into the local apparatus (passive acquisition)" such as waiting in a state where data transmitted from the external apparatus can be received, and based on a user input or an instruction based on a program, receiving data transmitted from the external apparatus, receiving data given from the external apparatus by distribution (or transmission, push notification, or the like), selecting and acquiring from the received data or information; and "new data is generated by editing data (conversion to text, rearrangement of data, extraction of partial data, change of file format, etc.) and the like, and the new data is acquired.

Returning to fig. 4 and 5, in S11, the analysis unit 12 analyzes the image of the transfer request form acquired by the acquisition unit 11 in S10, and identifies a value filled in each of a plurality of filled-in fields included in the transfer request form. The analysis unit 12 recognizes the contents filled in the transfer request form (image) through OCR reading and converts the contents into digital character codes.

For example, the analysis unit 12 identifies a numerical value filled in a plurality of first fill fields (the "amount" field shown in fig. 6) in which the transfer amount is filled, and a numerical value filled in a second fill field (the "subtotal field" shown in fig. 6) in which the sum of the numerical values filled in the plurality of first fill fields is filled. Note that the analysis unit 12 may also identify each value (including numbers, characters, symbols, etc.) in the other fill-in fields shown in fig. 6.

Returning to fig. 4 and 5, in S12, the determination unit 13 determines whether the recognition result of the analysis unit 12 in S11 satisfies an error condition. The error condition according to the present exemplary embodiment includes "the sum of the identification results of the values filled in the plurality of first filled-in fields and the identification result of the value filled in the second filled-in field do not match each other". The determination unit 13 calculates the total of the recognition results of the numerical values filled in the plurality of first filled-in fields. Then, the determination unit 13 determines whether the calculation result (the sum of the identification results of the numerical values filled in the plurality of first filled fields) and the identification result of the numerical value filled in the second filled field match each other. Note that another error condition may be further set.

When the recognition result of the analysis unit 12 satisfies at least one error condition (yes at S13), the processing unit 14 performs error processing (S15). For example, when the sum of the identification results of the numerical values filled in the plurality of first filled fields and the identification result of the numerical value filled in the second filled field do not match each other, the processing unit 14 performs error processing.

Error handling is any process for performing a manual validation job. For example, the processing unit 14 registers the image acquired by the acquisition unit 11 in S10, the recognition result of the analysis unit 12 in S11, the information determined to satisfy the error condition in S12, and the like in association with each other in a predetermined database (database of transfer request information satisfying the error condition). For example, the operator takes out the transfer request information registered on the database in a predetermined order by operating the terminal device, and displays the transfer request information on the display. Then, after the confirmation work, the correction work, and the like of the recognition result are performed via the terminal device, the operator inputs the recognition result (transfer request content) after the correction to the settlement host 20.

On the other hand, when the recognition result of the analysis unit 12 does not satisfy any error condition (no at S13), the processing unit 14 inputs the recognition result (transfer request content) of the analysis unit 12 to the settlement host 20 (S14).

The settlement host 20 performs a transfer process based on the input transfer request contents.

The processing apparatus 10 according to the present exemplary embodiment described above detects an erroneous recognition by OCR by determining whether or not an aggregate of recognition results of values filled in a plurality of first fill-in fields in which numerical values are filled and a recognition result of a value filled in a second fill-in field in which the aggregate of numerical values filled in the plurality of first fill-in fields match each other. By various feature processes, detecting the erroneous recognition of the OCR improves the detection accuracy, and the manual confirmation work can be reduced.

< second exemplary embodiment >

The processing apparatus 10 according to the present exemplary embodiment differs from the processing apparatus according to the first exemplary embodiment in the details of the error condition. The description is given in detail below.

The transfer form according to the present exemplary embodiment includes a plurality of third fill-in fields in which the same value is filled. In the case of the example shown in fig. 6, the subtotal field and the subtotal-mnemonic field correspond to the third padding field. The subtotal field is a field where the transfer requester fills in value, and the subtotal field is a field where the staff (staff of the financial institution) fills in value.

The analysis unit 12 recognizes the contents filled in the transfer request form (image) through OCR reading, and recognizes the values filled in the plurality of third filled fields (S11 in fig. 5). Note that the analysis unit 12 may also identify each value (including numbers, characters, symbols, etc.) in the other fill-in fields shown in fig. 6.

Then, the error condition according to the present exemplary embodiment includes "the recognition results of the values filled in the plurality of third filling fields do not match each other". The determining unit 13 determines whether the recognition result of the analyzing unit 12 satisfies the error condition in S11 (S12 in fig. 5). Note that another error condition (for example, the error condition described in the first exemplary embodiment) may be further set.

When the recognition result of the analysis unit 12 satisfies at least one error condition (yes at S13 in fig. 5), the processing unit 14 performs error processing (S15 in fig. 5). For example, when the identification results of the values filled in the plurality of third fill-in fields do not match each other, the processing unit 14 performs error processing. The details of the error processing are similar to those according to the first exemplary embodiment.

Note that other components of the processing apparatus 10 are similar to those according to the first exemplary embodiment.

The processing apparatus 10 according to the present exemplary embodiment described above detects an erroneous recognition of the OCR by determining whether recognition results of values filled in a plurality of third fill-in fields into which the same value is filled match each other. By various feature processes, detecting the erroneous recognition of the OCR improves the detection accuracy, and the manual confirmation work can be reduced. For example, by setting the error conditions described in the first and second exemplary embodiments, the detection accuracy of the erroneous recognition by the OCR is further improved.

< third exemplary embodiment >

The processing apparatus 10 according to the present exemplary embodiment differs from the processing apparatus according to each of the first and second exemplary embodiments in the details of the error condition. The description is given in detail below.

As shown in the example shown in fig. 6, the transfer form according to the present exemplary embodiment includes an item fill-in field in which items are filled in, and a specified transfer date fill-in field in which a specified transfer date is filled in.

The analysis unit 12 recognizes the contents filled in the transfer request form (image) through OCR reading, and recognizes the value filled in the item filling field and the value filled in the specified transfer date filling field (S11 of fig. 5). Note that the analysis unit 12 may also identify each value (including numbers, characters, symbols, etc.) in the other fill-in fields shown in fig. 6.

Then, the error condition according to the present exemplary embodiment includes "the condition that the recognition result of the values filled in the item fill-in field and the specified transfer date fill-in field does not satisfy the specifiable specified transfer date (specified date condition) determined for each item".

The determination unit 13 determines whether the identification result of the value filled in the item fill-in field and the identification result of the value filled in the specified transfer date fill-in field satisfy the specified date condition based on the condition information indicating the specified date condition as shown in fig. 7 (S12 in fig. 5).

Note that another error condition (e.g., the error condition described in the first and/or second exemplary embodiment) may be further set. Further, the details of one kind of item and the specified date condition illustrated herein are only one example, and the present invention is not limited thereto.

When the recognition result of the analysis unit 12 satisfies at least one error condition (yes at S13 in fig. 5), the processing unit 14 performs error processing (S15 in fig. 5). For example, when the identification result of the value filled in the item fill field and the identification result of the value filled in the specified transfer date fill field do not satisfy the specified date condition, the processing unit 14 performs error processing. Details of the error processing are similar to those according to the first exemplary embodiment.

Note that other components of the processing apparatus 10 are similar to those according to the first or second exemplary embodiment.

The processing device 10 according to the present exemplary embodiment described above detects an erroneous recognition of OCR by determining whether or not the recognition result of the value filled in the item fill-in field and the recognition result of the value filled in the specified transfer date fill-in field satisfy the condition (specified date condition) of the specifiable specified transfer date determined for each item. By various feature processes, detecting the erroneous recognition of the OCR improves the detection accuracy and can reduce the manual confirmation work. For example, by setting a plurality of error conditions described in the first, second, and third exemplary embodiments, the detection accuracy of the error recognition by the OCR is further improved.

< fourth exemplary embodiment >

The processing apparatus 10 according to the present exemplary embodiment differs from the processing apparatus according to each of the first, second, and third exemplary embodiments in the details of the error condition. The description is given in detail below.

As in the example shown in fig. 6, the transfer form according to the present exemplary embodiment includes a date fill-in field (specified transfer date fill-in field) in which a date (specified transfer date) is filled in, and a day fill-in field of the day of the week in which the specified transfer date is filled in.

The analysis unit 12 recognizes the contents filled in the transfer request form (image) through OCR reading, and recognizes the value filled in the date filled field and the value filled in the day filled field (S11 in fig. 5). Note that the analysis unit 12 may also identify each value (including numbers, characters, symbols, etc.) in the other fill-in fields shown in fig. 6.

Then, the error condition according to the present exemplary embodiment includes "the recognition result of the value filled in the date filling field and the recognition result of the value filled in the day filling field are contradictory".

The determination unit 13 determines whether the identification result of the value filled in the date filling field and the identification result of the value filled in the day filling field contradict based on the calendar information indicating the date and the day (S12 in fig. 5). "contradiction" herein means that the association between the date and the day of the week indicated by the calendar information is not matched.

Note that another error condition (for example, at least one of the error conditions described in the first, second, and third exemplary embodiments) may be further set.

When the recognition result of the analysis unit 12 satisfies at least one error condition (yes at S13 in fig. 5), the processing unit 14 performs error processing (S15 in fig. 5). For example, when the identification result of the value filled in the date filling field contradicts the identification result of the value filled in the day filling field, the processing unit 14 performs error processing. The details of the error processing are similar to those according to the first exemplary embodiment.

Note that other components of the processing apparatus 10 are similar to those according to the first, second, or third exemplary embodiments.

The processing apparatus 10 according to the present exemplary embodiment described above detects erroneous recognition by OCR by determining whether the recognition result of the value filled in the date filled-in field contradicts the recognition result of the value filled in the day filled-in field. By various feature processes, detecting the erroneous recognition of the OCR improves the detection accuracy and can reduce the manual confirmation work. For example, by setting a plurality of error conditions described in the first, second, third, and fourth exemplary embodiments, the detection accuracy of the erroneous recognition by the OCR is further improved.

< fifth exemplary embodiment >

The processing apparatus 10 according to the present exemplary embodiment differs from the processing apparatus according to each of the first, second, third, and fourth exemplary embodiments in the details of the error condition. The description is given in detail below.

The analysis unit 12 analyzes the image using a plurality of image analysis engines, and identifies values filled in a plurality of fill-in fields included in the transfer form (fill-in form) (S11 in fig. 5). The plurality of image analysis engines are different from each other in algorithm, setting, and the like of image analysis.

Then, the error condition according to the present exemplary embodiment includes "the recognition results of the plurality of image analysis engines do not match each other". The determination unit 13 determines whether the recognition results of the plurality of image analysis engines match each other (S12 in fig. 5). Note that another error condition (for example, at least one of the error conditions described in the first, second, third, and fourth exemplary embodiments) may be further set.

When the recognition result of the analysis unit 12 satisfies at least one error condition (yes at S13 in fig. 5), the processing unit 14 performs error processing (S15 in fig. 5). For example, when the recognition results of the plurality of image analysis engines do not match each other, the processing unit 14 performs error processing. The details of the error processing are similar to those according to the first exemplary embodiment.

Note that other components of the processing apparatus 10 are similar to those according to the first, second, third, or fourth exemplary embodiments.

The processing apparatus 10 according to the present exemplary embodiment described above detects an erroneous recognition by OCR by determining whether recognition results of a plurality of image analysis engines match each other. By various feature processes, detecting the erroneous recognition of the OCR improves the detection accuracy and can reduce the manual confirmation work. For example, by setting a plurality of error conditions described in the first, second, third, fourth, and fifth exemplary embodiments, the detection accuracy of the erroneous recognition by the OCR is further improved.

< sixth exemplary embodiment >

One example of a functional block diagram of the processing apparatus 10 according to the present exemplary embodiment is illustrated in fig. 8. As shown, the processing device 10 includes an acquisition unit 11, an analysis unit 12, a determination unit 13, a processing unit 14, a search unit 15, and a storage unit 16. The flowchart in fig. 9 shows one example of the flow of processing in the processing apparatus 10 according to the present exemplary embodiment.

First, in S20, the acquisition unit 11 acquires an image of a transfer form (fill-in form) generated by an image scanner or the like. The analysis unit 12 analyzes the image of the transfer request form acquired by the acquisition unit 11 in S20 using two image analysis engines (may be three or more image analysis engines) and identifies a value filled in each of a plurality of filled-in fields included in the transfer request form (S21 and S23).

Next, the search unit 15 searches the storage unit 16 using the recognition result of the value filled in the keyword filling field of each of the two image analysis engines as a keyword (S22 and S24).

As shown in fig. 10, the storage unit 16 stores a value (registration information) in a fill-in field of the transfer request form in association with the customer identification information (shown as "customer number"). For example, a transfer requester or the like who regularly transfers money to the same party registers information of the party in association with his/her own customer identification information. As information to be registered, at least some of a financial institution name, branch name, item, account number, name, and the like are exemplified.

In the case of the example shown in fig. 6, the keyword fill-in field is a field in which a customer number is filled. In other words, the search unit 15 searches the storage unit 16 using the identification result of the field in which the customer number is filled as a key. Then, when a keyword is found in the storage unit 16, the search unit 15 takes out the registration information associated with the keyword, and updates the recognition result of the analysis unit 12. Specifically, some of the recognition results of a plurality of fill-in fields (at least some of financial institution names, branch names, items, account numbers, names, and the like) included in the transfer request form are updated to previously registered information.

Next, when at least any one of the recognition results of the values respectively filled in the keyword filling fields of each of the two image analysis engines cannot be found in the storage unit 16 (no at S25), the processing unit 14 performs error processing (S32).

On the other hand, when each of the recognition results of the values filled respectively in the key filling fields of each of the two image analysis engines is found in the storage unit 16 (yes at S25), the determination unit 13 determines whether the recognition result of each of the two image analysis engines satisfies an error condition (S26 and S27). Herein, it is determined whether at least one of the error conditions described in the first to fourth exemplary embodiments is satisfied.

When any of the error conditions is satisfied (yes at S28), the processing unit 14 performs error processing (S32).

On the other hand, when any of the error conditions is not satisfied (no at S28), the determination unit 13 collates the recognition results of the two image analysis engines for the values of the plurality of filled-in fields, and determines whether the recognition results match each other (S29).

When the recognition results do not match (no at S30), the processing unit 14 performs error processing (S32).

On the other hand, when the recognition results match (yes at S30), the processing unit 14 inputs a transfer request generated based on the recognition result of the analyzing unit 12 and the registration information acquired from the storage unit 16 by the searching unit 15 to the settlement host 20 (S31).

Note that other components of the processing apparatus 10 are similar to those according to the first, second, third, fourth, or fifth exemplary embodiments.

The processing apparatus 10 according to the present exemplary embodiment described above improves the detection accuracy by detecting the erroneous recognition of the OCR through various kinds of feature processing. This enables the settlement host 20 to perform transfer processing without manual confirmation work.

< modification >

In this document, modified examples applicable to all the exemplary embodiments are described. Although it has been assumed so far that the fill form is a transfer request form, the fill form may be another form. The processing device 10 may then be utilized outside of the financial institution.

Some or all of the above exemplary embodiments may also be described as, but are not limited to, the following supplementary notes.

1. A processing apparatus, comprising:

an acquisition device for acquiring an image of a fill-in form including a plurality of first fill-in fields to which numerical values are filled and a second fill-in field to which a total of the numerical values filled in the plurality of first fill-in fields is filled;

analyzing means for analyzing the image and identifying a value filled in the plurality of first filled fields and a value filled in the second filled field;

determining means for determining whether or not the sum of the recognition results of the values filled in the plurality of first filled-in fields and the recognition result of the value filled in the second filled-in field match each other; and

processing means for performing error processing when the sum of the identification results of the values filled in the plurality of first filled-in fields and the identification result of the value filled in the second filled-in field do not match each other.

2. The processing apparatus according to supplementary note 1, wherein,

the fill-in form is a transfer request form, an

The first fill-in field is a fill-in field to fill in the transfer amount.

3. The processing apparatus according to supplementary note 1 or 2, wherein,

the fill-in form includes a plurality of third fill-in fields that fill in the same value,

the analyzing means analyzes the image and identifies a value filled in each of the plurality of third filling fields,

the determination means determines whether the recognition results of the values filled in the plurality of third filled-in fields match each other, and

when the recognition results of the values filled in the plurality of third fill-in fields do not match each other, the processing means performs error processing.

4. The processing apparatus according to supplementary note 3, wherein,

the fill-in form is a transfer request form, an

The third plurality of fill-in fields includes a fill-in field for a transfer requester fill-in value and a fill-in field for a staff fill-in value.

5. The processing apparatus according to any one of supplementary notes 1 to 4, wherein,

the fill-in form is a transfer request form, and includes an item fill-in field for filling in items and a specified transfer date fill-in field for filling in a specified transfer date,

the analyzing means analyzes the image and identifies the value filled in the item fill field and the value filled in the specified transfer date fill field,

the determination means determines whether the result of identification of the value filled in the item fill field and the result of identification of the value filled in the specified transfer date fill field satisfy the condition based on the condition information that determines the condition of the specifiable specified transfer date for each of the items, and

the processing means performs error processing when the identification result of the value filled in the entry fill field and the identification result of the value filled in the specified transfer date fill field do not satisfy the condition.

6. The processing apparatus according to any one of supplementary notes 1 to 5, wherein,

the fill form includes a date fill field to fill in the date and a day fill field to fill in the day,

the analyzing means analyzes the image and identifies the value filled in the date filling field and the value filled in the day filling field,

the determination means determines whether or not the identification result of the value filled in the date filling field and the identification result of the value filled in the day filling field contradict each other based on the calendar information indicating the date and the day of the week, and

the processing means performs error processing when the identification result of the value filled in the date fill-in field contradicts the identification result of the value filled in the day fill-in field.

7. The processing apparatus according to supplementary note 6, wherein,

the fill-in form is a transfer request form, an

The date fill-in field is a fill-in field that fills in the specified transfer date.

8. The processing apparatus according to any one of supplementary notes 1 to 7, wherein,

the analysis means analyzes the image using a plurality of image analysis engines, and identifies values filled in a plurality of fill-in fields included in the fill-in form, an

When the recognition results of the plurality of image analysis engines do not match each other, the processing means performs error processing.

9. A processing apparatus, comprising:

an acquisition means for acquiring an image of a transfer request form including a plurality of fill-in fields;

analyzing means for identifying values in a plurality of filled-in fields included in the transfer request form using a plurality of image analysis engines;

search means for searching the storage unit using, as a keyword, a recognition result of a value filled in a keyword filling field of each of the plurality of image analysis engines;

determining means for determining whether there is a value satisfying an error condition in recognition results of the values in the plurality of filled-in fields by the plurality of image analysis engines; and

processing means for performing error processing when the recognition results of the plurality of image analysis engines do not match each other, when a keyword generated by any one of the plurality of image analysis engines cannot be found in the storage unit, or when a value satisfying an error condition exists in the recognition results of the plurality of image analysis engines for values in the plurality of fill-in fields.

10. A method of processing, comprising:

by means of the computer, it is possible to,

acquiring an image of a filling form, wherein the filling form comprises a plurality of first filling fields filled with numerical values and a second filling field filled with the total numerical values filled in the plurality of first filling fields;

analyzing the image and identifying values filled in the plurality of first filled fields and values filled in the second filled fields;

determining whether an aggregate of the recognition results of the values filled in the plurality of first filled-in fields and the recognition result of the value filled in the second filled-in field match each other; and

when the sum of the recognition results of the values filled in the plurality of first filled-in fields and the recognition result of the value filled in the second filled-in field do not match each other, error processing is performed.

11. A program that causes a computer to function as:

an acquisition device for acquiring an image of a fill-in form including a plurality of first fill-in fields to which numerical values are filled and a second fill-in field to which a total of the numerical values filled in the plurality of first fill-in fields is filled;

analyzing means for analyzing the image and identifying a value filled in the plurality of first filled-in fields and a value filled in the second filled-in field;

determining means for determining whether or not the sum of the recognition results of the values filled in the plurality of first filled-in fields and the recognition result of the value filled in the second filled-in field match each other; and

processing means for performing error processing when the sum of the identification results of the values filled in the plurality of first filled-in fields and the identification result of the value filled in the second filled-in field do not match each other.

Although the invention of the present application has been described above with reference to the exemplary embodiments (and examples), the invention of the present application is not limited to the above-described exemplary embodiments (and examples). Various modifications, which will be apparent to those skilled in the art, may be made in the arrangement and details of the invention of the present application within the scope thereof.

The present application is based on and claims priority from japanese patent application No.2019-082587, filed 24/4/2019, the disclosure of which is incorporated herein by reference in its entirety.