JP6575158B2 - Information processing apparatus and information processing program - Google Patents

Description

  The present invention relates to an information processing apparatus and an information processing program.

  Japanese Patent Laid-Open Publication No. 2003-259542 has an object to correct an error in a character recognition function by using a character string notation and a notation rule, and an image input unit that inputs a numeric character string image, and a character image cutout that cuts out a character image. A character recognition unit that gives a character code, a recognition result storage unit that stores a recognition result, a notation analysis unit that analyzes a notation rule of a numeric character string to be read, and a numeric character stored in a numeric character string database A numeric character string data access unit that accesses the column data, a numeric character string database that stores the numeric character string data, a numeric character string collation unit that compares the recognition result with the numeric character string data stored in the database, and Numeric character string data selection section that selects the optimal numeric character string from the combined numeric character strings, and recognition result correction that corrects recognition errors and changes the attributes of recognition results An attribute change unit, be configured with a processing result output unit that outputs the processing results is disclosed.

  In Patent Document 2, an object is to provide an optical character reader that can more efficiently perform a correction operation of a number included in an arithmetic expression, and an image image of a form captured by a scanner is: It is handed over to the character recognition unit of the personal computer, character recognition is performed by the character recognition unit, and the recognition data is output. When the recognition result check / correction unit performs an operation check on this recognition data, Next, the fields in all the arithmetic expressions in which arithmetic errors are detected are considered as candidates, and for example, efficient movement is performed so that the fields are moved from the upper left to the lower right on the form. It is disclosed that the correction work proceeds in order.

  In Patent Document 3, when a key code recorded on a reading sheet is input by OCR, erroneous input due to misreading is reduced, and as a result, a reading rate for inputting a normal key code is improved. The key code is printed at a plurality of locations on the reading sheet, each key code is recognized by OCR, and the read data of the normal key code is determined using each recognition result corresponding to each key code. A character reading system provided with a computer system. The computer system, for example, determines whether the recognition result of each key code is correct or not based on a check digit added to each key code, and further verifies the normal recognition result. If there is a match, confirm one of the recognition results as a normal key code. Set the priority, to ascertain a high recognition result priority as a normal key code is disclosed.

Japanese Patent Application Laid-Open No. 07-114622 JP 2002-056354 A Japanese Patent Laid-Open No. 11-096285

The present invention relates to an information processing apparatus and information in which an object to be recognized is attached with an error correction code and the number of times of manual correction is reduced compared to the case where the recognition result is always manually corrected when an error is detected. The purpose is to provide a processing program.

The gist of the present invention for achieving the object lies in the inventions of the following items.
The invention of claim 1 relates to one recognition result among a plurality of recognition results for a recognition target with an error correction code , on condition that the recognition accuracy of the error correction code is higher or higher than a predetermined threshold. When an error is detected by the detection unit using an error correction code and an error is detected by the detection unit and a predetermined condition is satisfied, among the plurality of recognition results, the detection unit Conversion means for converting to a recognition result different from the target recognition result; first control means for controlling the detection result converted by the conversion means to perform detection by the detection means; A second control unit for controlling the recognition result of the recognition target to be manually corrected when a predetermined condition in the conversion unit is not satisfied; It is an information processing apparatus according to symptoms.

  According to a second aspect of the present invention, the predetermined condition is that the recognition accuracy of the recognition result of the conversion target by the conversion unit is higher or higher than the predetermined recognition accuracy, and the number of conversions by the conversion unit is The information processing apparatus according to claim 1, wherein the information processing apparatus is less than or less than a predetermined number of times or a combination thereof.

  The invention of claim 3 is characterized in that the control by the second control means is performed after the processing by the detection means, the conversion means, and the first control means is repeated a plurality of times. Item 3. The information processing device according to Item 1 or 2.

The invention according to claim 4 relates to one recognition result among a plurality of recognition results for a recognition target with an error correction code on condition that the recognition accuracy of the error correction code is higher or higher than a predetermined threshold. A correction unit that corrects an error of the recognition result using an error correction code; and when the error is corrected by the correction unit, the recognition result of the recognition target is determined based on information on the corrected recognition result. An information processing apparatus comprising control means for performing control so as to be corrected manually.

  The invention according to claim 5 is characterized in that the control means ranks in descending order according to the recognition accuracy of the corrected recognition result, and the rank of the recognition result is lower or lower than a predetermined rank, the correction When the recognition accuracy of the recognized recognition result is lower or lower than a predetermined recognition accuracy, or when a combination thereof is satisfied, the recognition result of the recognition target is controlled to be manually corrected. The information processing apparatus according to claim 4.

According to the invention of claim 6, the computer is provided with one of a plurality of recognition results for a recognition target with an error correction code on condition that the recognition accuracy of the error correction code is higher or higher than a predetermined threshold . A detection unit that performs error detection of the recognition result using an error correction code for the recognition result, and when an error is detected by the detection unit and satisfies a predetermined condition, among the plurality of recognition results, A conversion unit that converts a recognition result that is different from a recognition result that is a target of the detection unit, and a first control that controls the recognition result converted by the conversion unit to be detected by the detection unit. And a second control means for controlling the recognition result of the recognition target to be manually corrected when a predetermined condition in the conversion means is not satisfied An information processing program to function with.

According to the seventh aspect of the present invention, there is provided a computer comprising: one of a plurality of recognition results for a recognition target with an error correction code , provided that the recognition accuracy of the error correction code is higher or higher than a predetermined threshold . A correction unit that corrects an error in the recognition result using an error correction code for the recognition result, and when the error is corrected by the correction unit, based on information on the corrected recognition result, the recognition target It is an information processing program for functioning as a control means for controlling the recognition result to be corrected manually.

According to the information processing apparatus of claim 1, the number of times of manual correction is reduced as compared with the case where the recognition target has an error correction code and an error is detected and the recognition result is always corrected manually. Can do.

  According to the information processing apparatus of claim 2, the predetermined condition is that the recognition accuracy of the recognition result of the conversion target by the conversion unit is higher or higher than the predetermined recognition accuracy, and the number of conversions by the conversion unit. Is less than or less than a predetermined number of times, or a combination thereof.

  According to the information processing apparatus of the third aspect, the control by the second control unit can be performed after the processing by the detection unit, the conversion unit, and the first control unit is repeated a plurality of times.

  According to the information processing apparatus of claim 4, the number of times that the recognition target is attached with an error correction code, and that when the error is corrected, the recognition result is manually checked or corrected compared with the case where the recognition result is manually checked or corrected. Can be reduced.

  According to the information processing apparatus of claim 5, ranking is performed in descending order according to the recognition accuracy of the corrected recognition result, and the recognition result is lower than or lower than a predetermined order, corrected recognition When the recognition accuracy of the result is lower or lower than a predetermined recognition accuracy, or when these combinations are satisfied, the recognition result of the recognition target can be controlled to be manually corrected.

According to the information processing program of claim 6, when the recognition target has an error correction code and the error is detected, the number of times of manual correction is reduced compared to the case where the recognition result is always corrected manually. Can do.

  According to the information processing program of claim 7, the number of times that the recognition target is attached with an error correction code, and that when the error is corrected, the recognition result is always manually checked or corrected compared to the case where the recognition result is manually checked or corrected. Can be reduced.

It is a conceptual module block diagram about the structural example of 1st Embodiment. It is explanatory drawing which shows the system configuration example using this Embodiment. It is a flowchart which shows the process example by 1st Embodiment. It is a conceptual module block diagram about the structural example of 2nd Embodiment. It is a flowchart which shows the process example by 2nd Embodiment. It is a conceptual module block diagram about the structural example of 3rd Embodiment. It is a flowchart which shows the process example by 3rd Embodiment. It is a notional module block diagram about the structural example of 4th Embodiment. It is a flowchart which shows the process example by 4th Embodiment. It is explanatory drawing which shows the example of a form using this Embodiment. It is explanatory drawing which shows the example of a form of a prior art. It is a block diagram which shows the hardware structural example of the computer which implement | achieves this Embodiment.

First, before describing the present embodiment, the premise or an information processing apparatus using the present embodiment will be described with reference to FIGS. This description is intended to facilitate understanding of the present embodiment.
FIG. 10 is an explanatory diagram showing a form example using the present embodiment. As shown in the example of FIG. 10, there is a task of inputting 1018 a character string in a form having a number entry column 1010 in which various numbers such as a certificate number are entered by a data processing device 1020.
In order to perform this input, a character recognition technique is used in addition to manual input (so-called key input). However, current character recognition technology cannot always guarantee a 100% recognition rate, and may include recognition errors.
In addition, the certificate number or the like may include some error detection character (error detection code) in order to maintain data accuracy. When a character string is recognized for these number strings with error detection characters, if no error is detected, the recognition result is input as data, and if an error is detected, the recognition result is manually corrected and the data is input.
FIG. 11 is an explanatory diagram showing an example of a conventional technique for performing this operation.
The information processing apparatus 1100 includes a character recognition module 1110, an error detection module 1120, and a manual correction module 1130.
The character recognition module 1110 is connected to the error detection module 1120 and receives a form image (numerical string image) 1108. The character recognition module 1110 recognizes a character string in the form image (number string image) 1108.
The error detection module 1120 is connected to the character recognition module 1110 and the manual correction module 1130. The error detection module 1120 uses the error detection character in the recognition result, which is the processing result of the character recognition module 1110, to determine whether there is no error detection 1122 or error detection 1124. If it is determined that there is no error detection 1122, the recognition result of the character recognition module 1110 is output as it is as the data output 1128.
The manual correction module 1130 is connected to the error detection module 1120. If the error detection module 1120 determines that the error is detected to be 1124, the manual correction module 1130 causes the input person to check the form image (numerical string image) 1108 and the recognition result, and corrects it manually. . The result corrected by the manual correction module 1130 is used as a data output 1138.
When the accuracy of the character recognition module 1110 is low, man-hours for inputting data are increased.

Hereinafter, examples of various preferred embodiments for realizing the present invention will be described with reference to the drawings.
<< First Embodiment >>
FIG. 1 is a conceptual module configuration diagram of a configuration example according to the first embodiment.
The module generally refers to components such as software (computer program) and hardware that can be logically separated. Therefore, the module in the present embodiment indicates not only a module in a computer program but also a module in a hardware configuration. Therefore, the present embodiment is a computer program for causing these modules to function (a program for causing a computer to execute each procedure, a program for causing a computer to function as each means, and a function for each computer. This also serves as an explanation of the program and system and method for realizing the above. However, for the sake of explanation, the words “store”, “store”, and equivalents thereof are used. However, when the embodiment is a computer program, these words are stored in a storage device or stored in memory. This means that control is performed so as to be stored in the apparatus. Modules may correspond to functions one-to-one, but in mounting, one module may be configured by one program, or a plurality of modules may be configured by one program, and conversely, one module May be composed of a plurality of programs. The plurality of modules may be executed by one computer, or one module may be executed by a plurality of computers in a distributed or parallel environment. Note that one module may include other modules. Hereinafter, “connection” is used not only for physical connection but also for logical connection (data exchange, instruction, reference relationship between data, etc.). “Predetermined” means that the process is determined before the target process, and not only before the process according to this embodiment starts but also after the process according to this embodiment starts. In addition, if it is before the target processing, it is used in accordance with the situation / state at that time or with the intention to be decided according to the situation / state up to that point. When there are a plurality of “predetermined values”, the values may be different from each other, or two or more values (of course, including all values) may be the same. In addition, the description having the meaning of “do B when it is A” is used in the meaning of “determine whether or not it is A and do B when it is judged as A”. However, the case where it is not necessary to determine whether or not A is excluded.
In addition, the system or device is configured by connecting a plurality of computers, hardware, devices, and the like by communication means such as a network (including one-to-one correspondence communication connection), etc. The case where it implement | achieves by etc. is also included. “Apparatus” and “system” are used as synonymous terms. Of course, the “system” does not include a social “mechanism” (social system) that is an artificial arrangement.
In addition, when performing a plurality of processes in each module or in each module, the target information is read from the storage device for each process, and the processing result is written to the storage device after performing the processing. is there. Therefore, description of reading from the storage device before processing and writing to the storage device after processing may be omitted. Here, the storage device may include a hard disk, a RAM (Random Access Memory), an external storage medium, a storage device via a communication line, a register in a CPU (Central Processing Unit), and the like.

The information processing apparatus 100 according to the first embodiment recognizes the recognition target 108, and as illustrated in the example of FIG. 1, a character recognition module 110, an error detection module 120, a recognition result conversion module 130, A manual correction module 140 is provided.
The recognition target 108 may be a character image or stroke information of a line constituting the character (so-called online character recognition recognition target). The character may be any character to which an error detection code can be added. Of course, the characters may include numbers, and the recognition target 608 may be a number string. In addition, English characters may be included, and the recognition target 108 may be an English character string. Of course, the recognition target 108 may be an alphanumeric string.
The recognition target 108 may be a handwritten character or a printed character. The recognition target 108 is a character string with an error detection code. The character string here includes one or more characters as data and an error detection code, and is two or more characters. For example, when a character string is composed of a plurality of blocks, an error detection code may be included for each block. In that case, a plurality of error detection codes are included in the recognition target 108. The error detection code is at a predetermined position (for example, the rightmost character of the character string) in the recognition target 108. The error detection code may be used for maintaining the accuracy of data, and for example, a check digit (C / D) or the like may be used. Further, an error correction code may be used as the error detection code (that is, the error detection code is a concept including an error correction code here).
For example, a form image is input as the recognition target 108. In addition, an image of a numeric string column (for example, a certificate number) in the form may be input. There is no particular need to cut the column.

  The character recognition module 110 is connected to the error detection module 120 and the recognition result conversion module 130 and receives the recognition target 108. The character recognition module 110 recognizes the recognition target 108 as characters. The character recognition technique here may use a conventional character recognition technique. However, the character recognition module 110 can output a plurality of recognition results (text data) for each character as a recognition result. The recognition accuracy is also output to each of the recognition results (text data). Here, it is assumed that the higher the recognition accuracy value of the recognition result (text data) is, the higher the probability is that it is correct text data corresponding to the recognition target 108.

The error detection module 120 is connected to the character recognition module 110 and the recognition result conversion module 130. The error detection module 120 uses the error detection code for one recognition result among a plurality of recognition results for a recognition target with an error detection code, which is a processing result of the character recognition module 110, and performs error detection of the recognition result. . An existing method is used for error detection. Specifically, error detection is performed by comparing the error detection code and other recognition results among the recognition results by the character recognition module 110. If no error is detected (OK 122), the recognition result is used as the data output 128. When an error is detected (NG129), the recognition result conversion module 130 is caused to perform processing. In the error detection, if any character in the character string of the recognition result can be detected up to an error, the character that has detected the error is notified to the recognition result conversion module 130.
As a condition for the error detection module 120 to perform error detection, the recognition accuracy of the error detection code may be higher or higher than a predetermined threshold value. That is, the error detection module 120 may perform error detection when the recognition accuracy of the error detection code is higher or higher than a predetermined threshold.
The first character string targeted by the error detection module 120 is the character string having the highest recognition accuracy of each character recognized by the character recognition module 110. The character string targeted by the second or later error detection module 120 is a character string obtained by converting any character by the recognition result conversion module 130.

The recognition result conversion module 130 is connected to the character recognition module 110, the error detection module 120, and the manual correction module 140. When an error is detected by the error detection module 120 and a predetermined condition is satisfied, the recognition result conversion module 130 recognizes a recognition result different from the recognition result targeted by the error detection module 120 among the plurality of recognition results. Convert to Then, control is performed to cause the error detection module 120 to detect the converted recognition result.
Specifically, first, characters to be converted are extracted. Here, when the error detection module 120 detects up to which character in the character string of the recognition result has an error, the character may be the character to be converted. Further, if the error detection module 120 detects that there is an error in the recognition result, but has not been able to detect which character in the character string of the recognition result has an error, each error in the recognition result The character with the lowest character recognition accuracy (the character most likely to be misrecognized) may be the character to be converted.
The predetermined condition is that (1) the recognition accuracy of the recognition result to be converted by the recognition result conversion module 130 is higher or higher than the predetermined recognition accuracy, and (2) the recognition result conversion module 130 The number of conversions is less than or less than a predetermined number, or (3) a combination thereof. If the recognition accuracy of the recognition result to be converted is less than or less than a predetermined recognition accuracy, it is possible that the conversion result may be incorrect from the character of the current recognition result (the character targeted for error detection). Because it is expensive. If the number of conversions by the recognition result conversion module 130 is greater than or equal to a predetermined number, the character of the current recognition result (character that is the target of error detection) may be erroneous even after conversion. Because it is higher. In addition, these combinations are the conditions (logical product of the conditions (1) and (2)) which are the conditions (1) and (2).
When these conditions are satisfied, the recognition result conversion module 130 converts the character to be converted into a recognition result different from the recognition result targeted by the error detection module 120 among the plurality of recognition results of the character. Then, the error detection module 120 is processed again as a recognition result. More specifically, the first error detection target (the recognition result itself by the character recognition module 110) is the recognition result A having the highest recognition accuracy among a plurality of recognition results for the character to be converted (one character). It is. Since an error is detected in the character of the recognition result A, the character is converted into the recognition result B having the next highest recognition accuracy. That is, the recognition result A is converted into the recognition result B having the next higher recognition accuracy than the recognition accuracy of the recognition result A that is the current conversion target.
Then, the error detection module 120 performs error detection again for the character string whose recognition result A is the recognition result B. Here, the error detection processing and the recognition result conversion processing in the error detection module 120 and the recognition result conversion module 130 are performed until the recognition accuracy of the recognized character to be converted is less than or less than a predetermined threshold or predetermined. It is repeated until the number of times reached. If the recognition accuracy of the character to be converted is less than or less than a predetermined threshold value, or if the repetition is performed more or more than a predetermined number of times, the recognition result is output to the manual correction module 140 at that time To do.

  The recognition result A is converted to the recognition result B on the condition that the recognition accuracy of the recognition result B is higher than or equal to a predetermined threshold value, and the error detection module 120 performs the process again. Also good. If this condition is not satisfied (when conversion is not performed), the error correction module 120 may be caused to perform processing instead of the error detection module 120 to perform processing again. That is, the recognition result A may be converted into the recognition result B on condition that the recognition accuracy of the recognition result A and the recognition result B is higher or higher than a predetermined threshold. This is because the character to be converted (recognition result B) is preferably higher than or higher than a predetermined threshold as a character to be converted.

The manual correction module 140 is connected to the recognition result conversion module 130. When the predetermined condition in the recognition result conversion module 130 is not satisfied, the manual correction module 140 performs control so that the recognition result of the recognition target 108 is manually corrected. Specifically, when the conversion process is not performed by the recognition result conversion module 130, the data is displayed and the data input manually is received so that the data is manually input. The correction result is used as a data output 148. The correction here may include a new input in addition to changing a part of the recognition result.
Specifically, the manual correction module 140 (1) when a predetermined condition in the first recognition result conversion module 130 is not satisfied (that is, when the error detection of the error detection module 120 is one time), (2) The error detection module 120 performs error detection on the recognition result converted by the recognition result conversion module 130 (that is, performs error detection a plurality of times for one recognition target 108), and converts the recognition result. If conversion cannot be performed without satisfying the predetermined condition in the module 130, manual input is performed.
Note that there may be one input person or two or more input persons.

  The control by the manual correction module 140 may be performed after the processes by the error detection module 120 and the recognition result conversion module 130 are repeated a plurality of times. Specifically, when processing is performed by the recognition result conversion module 130 (when an error is detected by the error detection module 120), the processing by the error detection module 120 and the recognition result conversion module 130 may be repeated. .

FIG. 2 is an explanatory diagram showing a system configuration example using the present embodiment.
In the example shown in FIG. 2A, an example in the case of realizing as a stand-alone is shown.
The recognition target input device 210 is connected to the character recognition device 220. The recognition target input device 210 is a scanner or the like when the recognition target 108 is a character image. When the recognition target 108 is stroke information, the device is a device that detects the movement of the pen (such as an electronic pen or a tablet that detects the pen trajectory).
The character recognition device 220 includes the information processing device 100 and is connected to the recognition target input device 210 and the data processing device 230. The character recognition device 220 includes the information processing device 100, and performs processing as a character recognition device on the recognition target from the recognition target input device 210.
The data processing device 230 is connected to the character recognition device 220. The data processing device 230 processes the recognition result by the character recognition device 220. For example, it may be a form processing device or a storage device that stores recognition results.

In the example illustrated in FIG. 2B, an example in which processing is performed via another apparatus and a communication line is illustrated.
The recognition target input device 210A, the recognition target input device 210B, the character recognition device 220, the data processing device 230A, the data processing device 230B, the input terminal 240A, and the input terminal 240B are connected to each other via a communication line 290. The communication line 290 may be wireless, wired, or a combination thereof, and may be, for example, the Internet or an intranet as a communication infrastructure.
In this example, under the control of the manual correction module 140, an instruction is given to the input terminal 240, and the input person 242 performs correction work. For example, the image of the recognition target 108 and the recognition result may be displayed on the display device of the input terminal 240 so that the input person 242 corrects the recognition result.

FIG. 3 is a flowchart illustrating a processing example according to the first exemplary embodiment. In this example, the conversion is performed when the recognition accuracy of the recognition result to be converted is equal to or higher than a predetermined recognition accuracy.
In step S302, the character recognition module 110 receives the recognition target 108.
In step S304, the character recognition module 110 recognizes the recognition target 108. The character recognition module 110 recognizes the recognition target 108 with the error detection code received in step S302, and outputs a plurality of recognition candidate characters and their recognition accuracy as recognition results.
In step S306, the error detection module 120 performs error detection processing on the recognition result (or the conversion result in step S312) output from the character recognition module 110.

In step S308, the error detection module 120 determines whether or not an error has been detected. If there is an error, the process proceeds to step S310, and otherwise the process ends (step S399). That is, if no error is detected by the error detection module 120 (OK 122), the recognition result is output as it is as the data output 128. If an error is detected (NG 129), the process proceeds to step S310, and an error detection notification is output to the recognition result conversion module 130.
In step S310, the recognition result conversion module 130 determines whether or not the recognition accuracy of the conversion target is greater than or equal to a threshold value. If it is greater than or equal to the threshold value, the process proceeds to step S312; otherwise, the process proceeds to step S314. That is, the recognition result conversion module 130 has previously determined a character having the lowest recognition accuracy among unconverted recognition results (characters that are the above-described conversion target characters and may be erroneous). It is determined whether or not the threshold value is exceeded. If it is equal to or greater than the threshold, the process proceeds to step S312. If it is less than the threshold, the process proceeds to step S314.

In step S312, the recognition result conversion module 130 converts the recognition result. The recognition result conversion module 130 converts the character with the lowest recognition accuracy among the recognition results into the character with the next highest recognition accuracy among the other recognition candidate characters. Then, the process proceeds to step S306, and the process from step S306 (error detection process) to step S312 (recognition result conversion) is repeated.
In step S314, the manual correction module 140 corrects the recognition result manually. That is, the manual correction module 140 manually corrects data for the recognition result (or conversion result) when the recognition accuracy of the recognized character to be converted is less than the threshold.

<< Second Embodiment >>
FIG. 4 is a conceptual module configuration diagram of a configuration example according to the second embodiment.
The information processing apparatus 400 includes a character recognition module 110, an error detection module 120, a recognition result conversion module 130, a manual correction module 140, and a result integration module 450. In addition, the same code | symbol is attached | subjected to the site | part of the same kind as the above-mentioned embodiment, and the overlapping description is abbreviate | omitted (hereinafter the same). The result integration module 450 is added to the first embodiment.
The character recognition module 110 is connected to the error detection module 120 and the recognition result conversion module 130, receives the recognition target 108, and passes the recognition result 412 to the error detection module 120 and the recognition accuracy 414 to the recognition result conversion module 130. For example, a form image is input as the recognition target 108. In addition, an image of a numeric string column (for example, a certificate number) in the form may be input. In particular, the column need not be cut off, but an example of input as in the example shown in FIG. 4 is shown.

The error detection module 120 is connected to the character recognition module 110, the recognition result conversion module 130, and the result integration module 450. The error detection module 120 receives the recognition result 412 from the character recognition module 110 and the conversion result 432 from the recognition result conversion module 130. The recognition result 426 is passed to the integration module 450.
The recognition result conversion module 130 is connected to the character recognition module 110, the error detection module 120, and the manual correction module 140. The recognition result conversion module 130 receives the recognition accuracy 414 from the character recognition module 110, and converts the conversion result 432 into the error detection module 120. The conversion result 434 is passed to the module 140.
The manual correction module 140 is connected to the recognition result conversion module 130 and the result integration module 450, receives the conversion result 434 from the recognition result conversion module 130, and passes the manual input result 442 to the result integration module 450.
The result integration module 450 is connected to the error detection module 120 and the manual correction module 140. The result integration module 450 receives the recognition result 426 from the error detection module 120 and the manual input result 442 from the manual correction module 140, and outputs output data 458. The result integration module 450 integrates the recognition result 426, the converted recognition result (conversion result 432), and the result of the manual correction module 140 (manual input result 442). The integration here is to associate the recognition result (text data) with the recognition target 108, and includes the case where either the recognition result 426 or the manual input result 442 is selected. The output of the result integration module 450 becomes output data 458 of the information processing apparatus 400. This output data 458 becomes the contents of data input (recognition result of the recognition target 108).

FIG. 5 is a flowchart illustrating a processing example according to the second exemplary embodiment. The processing from step S502 to step S514 is equivalent to the processing from step S302 to step S314 in the flowchart shown in the example of FIG.
In step S502, the character recognition module 110 receives the recognition target 108.
In step S504, the character recognition module 110 recognizes the recognition target 108.
In step S506, the error detection module 120 performs an error detection process on the recognition (conversion) result.
In step S508, the error detection module 120 determines whether or not an error has been detected. If there is an error, the process proceeds to step S510. Otherwise, the process proceeds to step S516.
In step S510, the error detection module 120 determines whether or not the conversion target recognition accuracy is equal to or higher than the threshold value. If the conversion target recognition accuracy is equal to or higher than the threshold value, the process proceeds to step S512, and otherwise the process proceeds to step S514.
In step S512, the recognition result conversion module 130 converts the recognition result.
In step S514, the recognition result conversion module 130 manually corrects the result.
In step S516, the result integration module 450 integrates the results. That is, the recognition result 426, the converted conversion result, or the correction result (manual input result 442) corrected by the manual correction module 140 is integrated, and the process ends (step S599).

<< Third Embodiment >>
FIG. 6 is a conceptual module configuration diagram of an exemplary configuration according to the third embodiment.
The information processing apparatus 600 includes a character recognition module 610, an error correction module 620, a correction result determination module 630, and a manual correction module 640.
The character recognition module 610 is connected to the error correction module 620 and the correction result determination module 630 and receives the recognition target 608. The character recognition module 610 is equivalent to the character recognition module 110 of the first embodiment. The recognition target 608 is also equivalent to the recognition target 108 of the first embodiment, and includes an error correction code instead of the error detection code.

The error correction module 620 is connected to the character recognition module 610 and the correction result determination module 630. The error correction module 620 corrects an error of the recognition result by using the error correction code for one recognition result among a plurality of recognition results for the recognition target with the error correction code. An existing method is used for error correction. An error correcting code (ECC) may be used for maintaining the accuracy of data. For example, an error correcting code (ECC) is obtained by multiplying a binary representation of an original character string by a generator matrix. May be added. In this case, the decimal notation of the error correction code (ECC) is the error correction code. The error correction process is performed, for example, by a method in which the recognition result is binary-coded and multiplied by a parity check matrix. However, the present invention is not limited to this method, and an existing error correction process that can specify the error correction and its position can be used. The error correction result is output to the correction result determination module 630.
In addition, as a condition for the error correction module 620 to perform error correction, it may be a condition that the recognition accuracy of the error correction code is higher or higher than a predetermined threshold.
As a result of processing by the error correction module 620, when no error is detected, an error may be detected and corrected. When the error is corrected, information on how to correct which character in the character string of the recognition result is also passed to the correction result determination module 630.

The correction result determination module 630 is connected to the character recognition module 610, the error correction module 620, and the manual correction module 640. When the error is corrected by the error correction module 620, the correction result determination module 630 performs control so that the recognition result of the recognition target 608 is manually corrected based on the information regarding the corrected recognition result.
Specifically, if no error is detected by the error correction module 620 (OK 632), the recognition result by the character recognition module 610 is set as the data output 638.
When the error correction module 620 corrects the error, the correction result determination module 630 ranks in descending order according to the recognition accuracy of the recognition result corrected by the error correction module 620, and the rank of the recognition result is It is lower than a predetermined rank (being lower than a predetermined rank) or below (same as a predetermined rank or lower than that rank) ( 2) When the recognition accuracy of the corrected recognition result is lower than or lower than the predetermined recognition accuracy, or (3) a combination thereof is satisfied (NG634), the recognition result of the recognition target is manually set. You may make it control so that it may correct. In addition, these combinations are the conditions (logical product of the condition (1) and the condition (2)) that are the condition (1) and the condition (2). If there is a high probability that an error-corrected character is erroneously recognized, it will be corrected manually. In other cases (when there is a high probability that the error-corrected character is not erroneously recognized, OK 632), the error-corrected result is used as the data output 638.
Specifically, the correction result determination module 630 determines whether or not to output the correction result as it is based on the recognition accuracy of the recognized character at the position where the error is corrected. More specifically, if the recognition accuracy of the corrected character is higher than a predetermined threshold value among a plurality of recognition candidate characters, or if the candidate rank is higher than a certain threshold value, the correction result is assumed to be reliable. Data output 638 is assumed. Otherwise (when the above conditions (1) to (3) are satisfied), it is determined that the error correction by the error correction module 620 is not reliable, and a rejection notification (NG 634) is output so that manual input is performed to the manual correction module 640. To do.

  The manual correction module 640 is connected to the correction result determination module 630. The manual correction module 640 manually corrects the recognition result of the recognition target 608. Specifically, when there is a rejection notification (NG 634) that the correction is unreliable, the data is displayed and the data input manually is received so that the data is manually input. The correction here may include a new input in addition to changing a part of the recognition result. Further, the number of input users may be one, or two or more.

FIG. 7 is a flowchart illustrating a processing example according to the third exemplary embodiment. This example is an example in which manual input is performed when the recognition accuracy of the recognition result to be converted is equal to or higher than a predetermined recognition accuracy.
In step S <b> 702, the character recognition module 610 receives the recognition target 608.
In step S704, the character recognition module 610 recognizes the recognition target 608. The character recognition module 610 recognizes the recognition target 608 with the error correction code received in step S702, and outputs a plurality of recognition candidate characters and their recognition accuracy as a recognition result.

In step S706, the error correction module 620 performs error correction processing on the recognition result.
In step S708, the correction result determination module 630 determines whether or not the recognition accuracy of the corrected character is greater than or equal to a threshold value. If it is greater than or equal to the threshold value, the process ends (step S799). Otherwise, step S710 is performed. Proceed to In the correction result determination module 630, when the recognition accuracy of the corrected character is higher than a predetermined threshold value among the plurality of recognition candidate characters (or when the candidate rank is higher than a certain threshold value), the correction is assumed to be reliable (OK 632). Outputs the correction result (including the case of no error correction). Otherwise, it is determined that the correction is not reliable (NG 634), and a rejection notification is output to the manual correction module 640 so as to perform manual input.
In step S710, the manual correction module 640 manually corrects the result. That is, the manual correction module 640 manually corrects data for the correction result obtained by the error correction module 620.

<< Fourth Embodiment >>
FIG. 8 is a conceptual module configuration diagram of a configuration example according to the fourth embodiment.
The information processing apparatus 800 includes a character recognition module 610, an error correction module 620, a correction result determination module 630, a manual correction module 640, and a result integration module 850. The result integration module 850 is added to the third embodiment.
The character recognition module 610 is connected to the error correction module 620 and the correction result determination module 630, receives the recognition target 608, and passes the recognition result 812 to the error correction module 620 and the recognition accuracy 814 to the correction result determination module 630. For example, a form image is input as the recognition target 608. In addition, an image of a numeric string column (for example, a certificate number) in the form may be input. Although the column need not be cut out in particular, an example of input as shown in the example shown in FIG. 8 is shown.

The error correction module 620 is connected to the character recognition module 610 and the correction result determination module 630, receives the recognition result 812 from the character recognition module 610, and passes the correction result 822 to the correction result determination module 630.
The correction result determination module 630 is connected to the character recognition module 610, the error correction module 620, the manual correction module 640, and the result integration module 850, and the recognition accuracy 814 from the character recognition module 610 and the correction result 822 from the error correction module 620. And the correction result 836 is passed to the result integration module 850.
The manual correction module 640 is connected to the correction result determination module 630 and the result integration module 850, receives the recognition target 608, and passes the manual input result 842 to the result integration module 850.
The result integration module 850 is connected to the correction result determination module 630 and the manual correction module 640, receives the correction result 836 from the correction result determination module 630, receives the manual input result 842 from the manual correction module 640, and outputs output data 858. To do. The result integration module 850 integrates the error-corrected recognition result (correction result 836 (including the recognition result 812 when there is no error)) and the result of the manual correction module 640 (manual input result 842). The integration here is to associate the recognition result (text data) with the recognition target 608, and includes the case where either the correction result 836 or the manual input result 842 is selected. The output of the result integration module 850 becomes output data 858 of the information processing apparatus 800. The main output data 858 is data input content (recognition result of the recognition target 608).

FIG. 9 is a flowchart illustrating a processing example according to the fourth exemplary embodiment. The processing from step S902 to step S910 is equivalent to the processing from step S702 to step S710 in the flowchart shown in the example of FIG.
In step S902, the character recognition module 610 receives the recognition target 608.
In step S904, the character recognition module 610 recognizes the recognition target 608.
In step S906, the error correction module 620 performs error correction processing on the recognition result.
In step S908, the correction result determination module 630 determines whether or not the recognition accuracy of the corrected character is greater than or equal to a threshold value. If it is greater than or equal to the threshold value, the process proceeds to step S912. Otherwise, the process proceeds to step S910.
In step S910, the manual correction module 640 manually corrects the result.
In step S912, the result integration module 850 integrates the results. That is, the correction result 836 or the correction result (manual input result 842) corrected by the manual correction module 140 is integrated, and the process ends (step S899).

  A hardware configuration example of the information processing apparatus according to the present embodiment will be described with reference to FIG. The configuration shown in FIG. 12 is configured by a personal computer (PC), for example, and shows a hardware configuration example including a data reading unit 1217 such as a scanner and a data output unit 1218 such as a printer.

  A CPU (Central Processing Unit) 1201 includes various modules described in the above-described embodiments, that is, a character recognition module 110, an error detection module 120, a recognition result conversion module 130, a manual correction module 140, a result integration module 450, a character This is a control unit that executes processing according to a computer program that describes the execution sequence of each module such as the recognition module 610, the error correction module 620, the correction result determination module 630, the manual correction module 640, and the result integration module 850.

  A ROM (Read Only Memory) 1202 stores programs used by the CPU 1201, calculation parameters, and the like. A RAM (Random Access Memory) 1203 stores programs used in the execution of the CPU 1201, parameters that change as appropriate during the execution, and the like. These are connected to each other by a host bus 1204 configured by a CPU bus or the like.

  The host bus 1204 is connected to an external bus 1206 such as a PCI (Peripheral Component Interconnect / Interface) bus via a bridge 1205.

  A keyboard 1208 and a pointing device 1209 such as a mouse are input devices operated by an operator. The display 1210 includes a liquid crystal display device or a CRT (Cathode Ray Tube), and displays various types of information as text or image information.

  An HDD (Hard Disk Drive) 1211 includes a hard disk (may be a flash memory or the like), drives the hard disk, and records or reproduces a program executed by the CPU 1201 and information. The hard disk includes a recognition target 108, a recognition result 412, a recognition accuracy 414, no error detection 422, an error detection 424, a recognition result 426, a conversion result 432, a conversion result 434, a manual input result 442, output data 458, and a recognition target 608. , Recognition result 812, recognition accuracy 814, correction result 822, correction character OK832, correction character NG834, correction result 836, manual input result 842, output data 858, and the like are stored. Further, various other data, various computer programs, and the like are stored.

  The drive 1212 reads data or a program recorded in a removable recording medium 1213 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, and the data or program is read out to the interface 1207 and the external bus 1206. , The bridge 1205, and the RAM 1203 connected via the host bus 1204. The removable recording medium 1213 can also be used as a data recording area similar to a hard disk.

  The connection port 1214 is a port for connecting the external connection device 1215 and has a connection unit such as USB, IEEE1394. The connection port 1214 is connected to the CPU 1201 and the like via the interface 1207, the external bus 1206, the bridge 1205, the host bus 1204, and the like. The communication unit 1216 is connected to a communication line and executes data communication processing with the outside. The data reading unit 1217 is a scanner, for example, and executes document reading processing. The data output unit 1218 is a printer, for example, and executes document data output processing.

  Note that the hardware configuration of the information processing apparatus illustrated in FIG. 12 illustrates one configuration example, and the present embodiment is not limited to the configuration illustrated in FIG. 12, and the modules described in the present embodiment are executed. Any configuration is possible. For example, some modules may be configured with dedicated hardware (for example, Application Specific Integrated Circuit (ASIC), etc.), and some modules are in an external system and connected via a communication line In addition, a plurality of systems shown in FIG. 12 may be connected to each other via communication lines so as to cooperate with each other. In particular, in addition to personal computers, portable information communication devices (including mobile phones, smartphones, mobile devices, wearable computers, etc.), information appliances, robots, copiers, fax machines, scanners, printers, multifunction devices (scanners, printers, An image processing apparatus having two or more functions such as a copying machine and a fax machine) may be incorporated.

In the above-described embodiment, an example in which a recognition target is a numeric string has been described. However, as long as an error detection code or an error correction code is attached, not only numbers but other character types (for example, alphabetic characters) Etc.) or a combination thereof (for example, an alphanumeric string or the like) may be the recognition target.
Further, in the description of the above-described embodiment, “more than”, “less than”, “greater than”, and “less than (less than)” in a comparison with a predetermined value contradicts the combination. As long as the above does not occur, “larger”, “smaller (less than)”, “more than”, and “less than” may be used.

The program described above may be provided by being stored in a recording medium, or the program may be provided by communication means. In that case, for example, the above-described program may be regarded as an invention of a “computer-readable recording medium recording the program”.
The “computer-readable recording medium on which a program is recorded” refers to a computer-readable recording medium on which a program is recorded, which is used for program installation, execution, program distribution, and the like.
The recording medium is, for example, a digital versatile disc (DVD), which is a standard established by the DVD Forum, such as “DVD-R, DVD-RW, DVD-RAM,” and DVD + RW. Standard “DVD + R, DVD + RW, etc.”, compact disc (CD), read-only memory (CD-ROM), CD recordable (CD-R), CD rewritable (CD-RW), Blu-ray disc ( Blu-ray (registered trademark) Disc), magneto-optical disk (MO), flexible disk (FD), magnetic tape, hard disk, read-only memory (ROM), electrically erasable and rewritable read-only memory (EEPROM (registered trademark)) )), Flash memory, Random access memory (RAM) SD (Secure Digital) memory card and the like.
The program or a part of the program may be recorded on the recording medium for storage or distribution. Also, by communication, for example, a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a wired network used for the Internet, an intranet, an extranet, or a wireless communication It may be transmitted using a transmission medium such as a network or a combination of these, or may be carried on a carrier wave.
Furthermore, the program may be a part of another program, or may be recorded on a recording medium together with a separate program. Moreover, it may be divided and recorded on a plurality of recording media. Further, it may be recorded in any manner as long as it can be restored, such as compression or encryption.

DESCRIPTION OF SYMBOLS 100 ... Information processing apparatus 110 ... Character recognition module 120 ... Error detection module 130 ... Recognition result conversion module 140 ... Manual correction module 210 ... Recognition target input device 220 ... Character recognition device 230 ... Data processing device 240 ... Input terminal 290 ... Communication Line 400 ... Information processing device 450 ... Result integration module 600 ... Information processing device 610 ... Character recognition module 620 ... Error correction module 630 ... Correction result determination module 640 ... Manual correction module 800 ... Information processing device 850 ... Result integration module

Claims (7)

With the proviso that recognition accuracy of the error correction code is preliminarily higher than the determined threshold or higher, by using an error correction code for one of the recognition results of the plurality of recognition results for the recognition target with error correction code, Detecting means for detecting an error in the recognition result;
When an error is detected by the detection unit and a predetermined condition is satisfied, a conversion unit that converts a recognition result that is different from the recognition result that is the target of the detection unit among the plurality of recognition results;
First control means for controlling the recognition result converted by the conversion means to perform detection by the detection means;
An information processing apparatus comprising: a second control unit configured to perform control so that the recognition result of the recognition target is manually corrected when a predetermined condition in the conversion unit is not satisfied. The predetermined condition is that the recognition accuracy of the recognition result of the conversion target by the conversion unit is higher or higher than the predetermined recognition accuracy, and the number of conversions by the conversion unit is less than the predetermined number of times. The information processing apparatus according to claim 1, wherein the information processing apparatus is the following or a combination thereof. 3. The control by the second control unit is performed after the processing by the detection unit, the conversion unit, and the first control unit is repeated a plurality of times. 4. Information processing device. Using the error correction code for one recognition result of a plurality of recognition results for a recognition target with an error correction code , on condition that the recognition accuracy of the error correction code is higher or higher than a predetermined threshold , Correction means for correcting an error in the recognition result;
Control means for controlling so that the recognition result of the recognition target is manually corrected based on the information related to the corrected recognition result when an error is corrected by the correction means. Processing equipment. The control means ranks in descending order according to the recognition accuracy of the corrected recognition result, and the recognition result rank is lower or lower than a predetermined rank, and the recognition accuracy of the corrected recognition result 5. The control is performed so that the recognition result of the recognition target is manually corrected when the recognition accuracy is lower than or less than a predetermined recognition accuracy, or a combination thereof is satisfied. Information processing device. Computer
With the proviso that recognition accuracy of the error correction code is preliminarily higher than the determined threshold or higher, by using an error correction code for one of the recognition results of the plurality of recognition results for the recognition target with error correction code, Detecting means for detecting an error in the recognition result;
When an error is detected by the detection unit and a predetermined condition is satisfied, a conversion unit that converts a recognition result that is different from the recognition result that is the target of the detection unit among the plurality of recognition results;
First control means for controlling the recognition result converted by the conversion means to perform detection by the detection means;
An information processing program for functioning as a second control unit that performs control so that the recognition result of the recognition target is manually corrected when a predetermined condition in the conversion unit is not satisfied. Computer
Using the error correction code for one recognition result of a plurality of recognition results for a recognition target with an error correction code , on condition that the recognition accuracy of the error correction code is higher or higher than a predetermined threshold , Correction means for correcting an error in the recognition result;
An information processing program for functioning as a control unit that controls to manually correct a recognition result of the recognition target based on information on the corrected recognition result when an error is corrected by the correction unit.

Images