JP4190159B2 - Character recognition processing system and program - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a character recognition processing system and program. More specifically, the present invention performs character recognition processing based on data entered in a form. For data that has not been character-recognized, the form has been subjected to character recognition processing in the past. The present invention also relates to a character recognition processing system and program capable of increasing the character recognition rate by estimating the contents based on the data.
[0002]
[Prior art]
Usually, in a system for inputting data entered in a form, it is often performed to perform character recognition using OCR for streamlining operations and simplify data input.
[0003]
In this type of character recognition processing system, repeatedly generated data is stored in the database as processing results, and when similar data is generated next time, past processing results are extracted from the stored database and used. Therefore, there are many cases in which data input is labor-saving. In this case, the data entered in the form is processed as follows.
[0004]
1) The form data recognized daily is updated and updated in the database as the processing results.
[0005]
2) When character recognition of new form data is performed, character recognition of the read form data is performed based on the image image of the form data by OCR. As a result of this character recognition processing, when there is data that has not been character-recognized, the following processing 3) and subsequent steps are performed.
[0006]
3) If there is data that cannot be recognized by the character entered in the form even if the character recognition processing is performed by OCR, the form data that matches the data of the key item that is a predetermined data item is obtained. Search from the database.
[0007]
For example, as shown in FIG. 6, in a transfer request form including items of bank name, branch name, subject, account number, payee name, transfer amount, and client name, character recognition of data entered in any item is performed. If there is data that cannot be found, the bank name, branch name, subject, and account number items are determined in advance as search keys, and a form that matches the data of these search keys is searched from the database.
[0008]
4) Then, as a result of searching from the database, if there is a form that matches the data of the item defined as the search key, the data on the correction screen is replaced with the data of the searched form.
[0009]
In the example shown in FIG. 6, as a result of character recognition processing, the recipient name was not correctly recognized as “Yama ?? Taro”, but in 3) the bank name, branch name, subject, and account number were set as search keys, As a result of searching the database for a form that matches the search key, “Yamada Taro” is the recipient name that matches this bank name “XX Ginkgo”, branch name “Hongten”, subject “ordinary”, and account number “12345”. It was searched. Therefore, the recipient name “Yama ?? Taro” whose character was not recognized is replaced with the data “Yamada Taro”.
[0010]
In this manner, by using a form and its data that have been subjected to character recognition processing in the past, support for character recognition of data that has not been character-recognized is provided.
[0011]
[Problems to be solved by the invention]
However, such a conventional character recognition processing system has the following problems.
[0012]
That is, in the character recognition processing system as described above, the item to be used as the search key is determined in advance, and once the search key is determined, the search key is changed for each form for character recognition processing. I can't do that.
[0013]
Therefore, the data for the item determined as the search key cannot be used when the characters are not correctly recognized. For example, as shown in FIG. 6, when performing character recognition processing of form data, each item of bank name, branch name, item, and account number is specified as a search key, and the data of all the specified items are correctly written. When it is recognized, it functions correctly as a search key, but cannot be used when there is a character that is not recognized even in one of the items specified as the search key.
[0014]
FIG. 7 is an example showing such a case, and shows a case where the data of the branch name among the items specified as the search key is not recognized.
[0015]
In this way, if there is unrecognized data in the item specified as the search key, it is not possible to use the previously processed form data stored in the database, so fill in the form. There is a problem that the operator must read the input data and manually input it directly, increasing the burden on the operator.
[0016]
The present invention has been made in view of such circumstances, and allows an item to be set as a search key to be arbitrarily set, and the estimation of data that has not been recognized in the character recognition processing can be effectively performed on data accumulated in the past. It is an object of the present invention to provide a character recognition processing system and program that can be utilized to increase the character recognition rate and reduce the burden on the operator.
[0017]
[Means for Solving the Problems]
In order to achieve the above object, the present invention takes the following measures.
[0018]
That is, according to the first aspect of the present invention, character recognition means for performing character recognition processing on each data of a form composed of a plurality of items based on the image image of the form, and a form that has been recognized in the past as each item of the form. If there is data that has not been character-recognized among the form data storage means stored together with the corresponding data and the data that has been character-recognized by the character recognition means, the character recognition means A selection means for selecting, based on the item, an item to be a search key for searching the form stored in the form data storage means and the data of each item from items corresponding to the character-recognized data; A form including data corresponding to the item of the search key selected by the selection means and data corresponding to each item are searched from the form data storage means. Searching means, estimation means for estimating data not recognized by the character based on the form searched by the searching means and data corresponding to each item, and data not recognized by the data estimated by the estimating means With replacement means to replace When there are a plurality of forms searched by the search means, the degree of coincidence between the data corresponding to each item of the searched form and the data corresponding to each item of the form subjected to character recognition processing by the character recognition means is calculated. The calculation means for calculating for each searched form is added, and the estimation means is a data that has not been recognized based on the data corresponding to each item of the form having the maximum degree of coincidence calculated by the calculation means. Estimate .
[0019]
Therefore, in the character recognition processing system according to the first aspect of the present invention, by taking the above-mentioned means, it is possible to designate any item as a search key from a form consisting of a plurality of items. By retrieving similar data from past data based on the search key, it is possible to estimate data that has not been recognized with high accuracy.
[0021]
Therefore, the claims 1 In the character recognition processing system of the invention, if there are a plurality of past forms searched based on the search key by taking the above-mentioned means, they are not recognized based on the form with higher matching degree. Since the estimated data can be estimated, the estimation accuracy can be further increased.
[0022]
Claim 2 In the invention of claim 1 In the character recognition processing system according to the invention, the data replaced by the replacing means is displayed using different display colors, different character sizes, or different modification methods, whereby each data of the form is displayed as an image of the form. A display means for displaying the change data and a change input means for receiving a change input of the replaced data displayed by the display means are added.
[0023]
Therefore, the claims 2 In the character recognition processing system according to the invention, by taking the above-described means, it is possible to easily recognize that the data has been replaced with respect to the replaced data. In addition, when the data needs to be changed by comparing with the input image, the data to be changed can be input.
[0024]
Claim 3 In the present invention, character recognition processing is performed based on data entered in a form corresponding to each item, and for data that has not been character-recognized, a form that has been character-recognized in the past and data corresponding to each item A computer-readable program for increasing the character recognition rate by estimating the content of the data based on the character recognition means for performing character recognition processing on each data of the form consisting of a plurality of items based on the image image of the form, There was data that was not character-recognized in the form data storage means that stored the form that was subjected to character recognition processing together with the data corresponding to each item of the form, and each data that was character-recognized by the character recognition means In the case, the book stored in the form data storage means is based on the data item recognized by the character recognition means. And a selection means for selecting an item to be a search key for searching the data of each item from items corresponding to the character-recognized data, and data corresponding to the item of the search key selected by the selection means. A retrieval unit that retrieves data corresponding to each of the included forms and the respective items from the form data storage unit, an estimation that estimates the data that has not been recognized based on the forms retrieved by the retrieval unit and the data corresponding to each of the items Replacement means for replacing data that has not been recognized by the data estimated by the estimation means When there are a plurality of forms searched by the search means, the degree of coincidence between the data corresponding to each item of the searched form and the data corresponding to each item of the form subjected to character recognition processing by the character recognition means Is calculated for each searched form, and the estimation means is configured to calculate the data that has not been recognized based on the data corresponding to each item of the form having the maximum matching degree calculated by the calculating means. Means to estimate Is a program that causes a computer to realize
[0025]
Therefore, the claims 3 In the program of the invention, it is possible to specify an arbitrary item as a search key from a plurality of items by taking the above-described means, and past data based on the specified search key By retrieving similar data from the above, it is possible to estimate data that has not been recognized with high accuracy.
[0026]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
[0027]
An embodiment of the present invention will be described with reference to FIGS.
[0028]
FIG. 1 is a configuration diagram showing an example of the overall configuration of a character recognition processing system according to an embodiment of the present invention.
[0029]
A character recognition processing system 1 according to an embodiment of the present invention includes a computer such as a workstation, and includes an image data conversion unit 2, an image image data storage database (image image data storage DB) 3, a character recognition unit 4, and the like. , Character recognition data storage database (character recognition data storage DB) 5, search key selection unit 6, form data search unit 7, coincidence calculation unit 8, character estimation unit 9, character replacement unit 10, replacement A post-character recognition data storage database (post-replacement character recognition data storage DB) 11, a display unit 12, a change input unit 13, and a form data storage database (form data storage DB) 14 are provided.
[0030]
The image data conversion unit 2 uses, for example, a scanner, converts data handwritten in a form into image data, stores the converted image data in the image image data storage DB 3, and outputs the image data to the character recognition unit 4. .
[0031]
The image image data storage DB 3 stores the image data of the form output from the image data conversion unit 2. Further, based on a request from the display unit 12, the stored image data is output to the display unit 12.
[0032]
The character recognition unit 4 uses, for example, OCR, and performs character recognition processing based on the image data output from the image data conversion unit 2. Then, the character recognition data as a result of character recognition is stored in the character recognition data storage DB 5 and is output to the search key selection unit 6. FIG. 2 shows an example of a form subjected to character recognition processing by the character recognition unit 4. This form consists of the following items: bank name, branch name, item, account number, payee name, transfer amount, and client name. The bank name, item, account number, transfer amount, and client name are correct characters. The case where the branch name and the recipient name are not recognized is recognized.
[0033]
The character recognition data storage DB 5 stores character recognition data output from the character recognition unit 4 as shown in FIG. Further, based on a request from the search key selection unit 6, the stored character recognition data is output to the search key selection unit 6.
[0034]
Based on the character recognition data output from the character recognition unit 4 or the character recognition data acquired as necessary from the character recognition data storage DB 5, the search key selection unit 6 performs character recognition on any of the items. If there is data that has not been processed, an item serving as a search key for searching for the form stored in the form data storage DB 14 and the data corresponding to each item based on the item whose data has been character-recognized. Then, the item corresponding to the character-recognized data is selected, and the selection result is output to the form data search unit 7.
[0035]
For example, in the case of the character recognition data as shown in FIG. 2, the branch name and the recipient name are items that are not recognized. This item is selected as a search key from among the items that have been character-recognized: bank name, subject, account number, transfer amount, and transfer person name. The number of items to be selected as the search key is arbitrary.
[0036]
On the other hand, the search key selection unit 6 recognizes all the data of each item based on the character recognition data output from the character recognition unit 4 or the character recognition data acquired as necessary from the character recognition data storage DB 5. If it is, the character recognition data is output to the display unit 12.
[0037]
The form data search unit 7 searches the form data storage DB 14 for a form including the search key item selected by the search key selection unit 6 and data corresponding to each item, and sends the search result to the matching degree calculation unit 8. Output. FIG. 3 shows an example of a form searched from the form data storage DB 14 and data corresponding to the item when a bank name, a subject, and an account number are selected as a search key in the search key selection unit 6. . As a result of searching with the search key “XX Ginkgo” as the bank name, “Normal” as the subject, “12345” as the account number, the data of “Shinjuk” as the branch name and “Sato Hanako” as the recipient name 1) and the branch name As “Honten” and “Yamada Taro” as the recipient name 2).
[0038]
The degree of coincidence calculation unit 8 calculates the degree of coincidence of data with the form subjected to character recognition processing by the character recognition unit 4 for the form data searched by the form data search unit 7. When there are a plurality of forms searched by the form data search unit 7, the degree of coincidence with each data of the form subjected to the character recognition processing by the character recognition unit 4 is calculated for each searched form. Then, the degree of coincidence as a result of the calculation is output to the character estimation unit 9.
[0039]
The degree of coincidence α is calculated based on, for example, the following expression (1).
That is,
α = Σf _n (R _n , J _n ) / Len (r _n , J _n ) × φ _n ... (1)
However,
Σf _n (X, y): A function for comparing the character string x and the character string y to obtain the number of matched characters.
len (x, y): A function for obtaining the value of the larger number of characters for the character string x and the character string y.
r _n : N-th item of recognition result.
j _n : The nth item of the result search result.
φ _n : Weight coefficient of the nth item.
[0040]
Φ _n The (weight coefficient of the nth item) is set to perform weighting in consideration of the recognition accuracy depending on the character type of each item. For example, a numerical item with high recognition accuracy is high and katakana with low recognition accuracy. Set Kanji items low.
[0041]
The character estimation unit 9 is a case where only one form is searched by the form data search unit 7, and the coincidence degree α made by the coincidence degree calculation unit 8 for the form is equal to or greater than a predetermined reference value. Is positive for the contents of the retrieved form. In this case, the character replacement unit 10 replaces the data not recognized by the character recognition unit 4 based on the form data searched by the form data search unit 7 and replaces the replaced data. Set the replacement flag so that can be identified. The replacement flag will be described later.
[0042]
Further, when a plurality of forms are searched by the form data search unit 7, the character estimation unit 9 sets the content of the form having the maximum coincidence α made by the coincidence calculation unit 8 as positive. In this case, the character replacement unit 10 replaces the data not recognized by the character recognition unit 4 based on the form data searched by the form data search unit 7 and replaces the replaced data. Set the replacement flag so that can be identified.
[0043]
FIG. 4 is a diagram showing a state in which data is replaced by the character replacement unit 10 and a replacement flag is set for the replaced data. When the degree of coincidence α with the form shown in FIG. 2 is calculated for each of data 1) and data 2) shown in FIG. 3, data 2) has a larger degree of coincidence α. Therefore, in the form shown in FIG. 2, the branch name is “honten” entered in data 2) is positive, and the recipient is also “Yamada Taro” entered in data 2). Therefore, the branch name “Hon ??” in the form of FIG. 2 is replaced with “Honten”, and the recipient name “Yama ?? Taro” is replaced with “Yamada Taro”. Further, for the replaced data, the replacement flag data attached with parentheses together with the data is “1”. The data of the bank name, subject, and account number in the form shown in FIG. 2 is not replaced, so the replacement flag data is “0”. Note that the transfer amount and the client name in the form data shown in FIG. 2 are not compared with the form data shown in FIG.
[0044]
As shown in FIG. 4, the character replacement unit 10 stores the replaced character recognition data after the data replacement in the post-replacement character recognition data storage DB 11 and outputs it to the display unit 12. Note that even when the form data is not searched by the form data search unit 7 or when there is the form data searched by the form data search unit 7, the matching degree α calculated by the matching degree calculation unit 8 is used. If the value of is less than the predetermined reference value, the data is not replaced. The character replacement unit 10 outputs the character recognition data output from the character estimation unit 9 to the display unit 12 even when the data replacement is not performed as described above.
[0045]
The post-replacement character recognition data storage DB 11 stores the output post-replacement character recognition data when the post-replacement character recognition data as shown in FIG. 4 is output from the character replacement unit 10. Further, based on a request from the display unit 12, the stored post-replacement character recognition data is output to the display unit 12.
[0046]
The display unit 12 displays the post-replacement character recognition data output from the character replacement unit 10 or the post-replacement character recognition data storage DB 11 together with the image data of this form stored in the image image data storage DB 3. In this case, the data replaced by the character replacement unit 10 is displayed using a different display color, a different character size, or a different modification method by recognizing that the replacement flag is “1”. This makes it possible to identify the replaced data. In order to clarify the accuracy of the replaced data, the colors may be displayed separately according to the value of the coincidence α.
[0047]
The display unit 12 includes the unrecognized data when the character replacement unit 10 outputs character recognition data that has not been recognized but has not been replaced. The character recognition data thus displayed is displayed on the screen together with the image data of this form stored in the image image data storage DB 3.
[0048]
Further, when the character recognition data in which all the data is recognized is output from the search key selection unit 6, the display unit 12 displays this character recognition data as an image of this form stored in the image image data storage DB3. Display on screen with data.
[0049]
In this way, the character recognition data and the image data corresponding to the form are displayed on the screen in parallel on the display unit 12 so that the operator can compare the contents of both data.
[0050]
As a result of the comparison between the two data, if the operator determines that the character recognition data includes erroneously recognized data or erroneously replaced data, the change input unit 13 inputs the change data. Thus, it is possible to change the contents of arbitrary character recognition data.
[0051]
When the character recognition data includes data that has not been character-recognized, the operator reads the data from the image data of the form corresponding to the character recognition data, and reads the data. Manual input from the change input unit 13 is possible.
[0052]
When the operator inputs change data or unrecognized data, the change input unit 13 updates the content of the character recognition data based on the input data.
[0053]
When all data input from the change input unit 13 is completed, the operator presses a confirmation key (not shown) provided in the change input unit 13 to confirm and confirm the character recognition data. The character recognition data is output to the form data storage DB 14.
[0054]
The form data storage DB 14 stores the character recognition data output from the change input unit 13 as a form for which character recognition processing has been performed in the past and data for each item of the form. The stored form and the data for each item of the form are used for the search performed by the form data search unit 7.
[0055]
The character recognition processing system 1 according to the embodiment of the present invention configured as described above is realized by a computer that reads a program stored in a storage medium such as a magnetic disk and whose operation is controlled by this program.
[0056]
Next, the operation of the character recognition processing system according to the embodiment of the present invention configured as described above will be described with reference to the flowchart shown in FIG.
[0057]
When processing data entered in a form using the character recognition processing system 1 according to the embodiment of the present invention, first, a form in which data is handwritten in each item is processed by the image data conversion unit 2. It is converted into image data (S1). The converted image data is output from the image data conversion unit 2 to the image image data storage DB 3, stored therein, and output to the character recognition unit 4.
[0058]
The image data output from the image data conversion unit 2 to the character recognition unit 4 is subjected to character recognition processing by the character recognition unit 4 (S2). Character recognition data as a result of character recognition is output from the character recognition unit 4 to the character recognition data storage DB 5, stored therein, and output to the search key selection unit 6.
[0059]
Character recognition data is output from the character recognition unit 4 to the search key selection unit 6. Moreover, it is also possible to acquire the character recognition data memorize | stored in character recognition data storage DB5 as needed. In the search key selection unit 6, no character was recognized in each data based on the character recognition data output from the character recognition unit 4 or the character recognition data acquired as necessary from the character recognition data storage DB 5. If there is data (S3: Yes), it becomes a search key for searching the form stored in the form data storage DB 14 and the data corresponding to each item based on the data-recognized data item. An item is selected from items corresponding to character-recognized data (S4). The search key selected in this way is output to the form data search unit 7.
[0060]
On the other hand, when all the data of each item is character-recognized based on the character-recognition data output from the character-recognition part 4, or the character-recognition data acquired as needed from the character-recognition data storage DB5 (S3: No) The character recognition data is output from the search key selection unit 6 to the display unit 12 (S11).
[0061]
Then, the form data search unit 7 searches the form data storage DB 14 for the form including the item of the search key and the data corresponding to each item, and the search result is output to the coincidence degree calculation unit 8 (S5).
[0062]
In the coincidence degree calculation unit 8, the degree of coincidence α of the data corresponding to each item of the form searched by the form data search unit 7 with the data corresponding to each item of the form subjected to character recognition processing by the character recognition unit 4. An operation is performed (S6). When there are a plurality of forms searched by the form data search unit 7, the matching degree α with the data corresponding to each item of the form subjected to character recognition processing by the character recognition unit 4 is determined for each searched form. Is calculated. Then, the degree of coincidence α, which is the calculated result, is output to the character estimation unit 9.
[0063]
In the character estimation unit 9, when only one form is searched by the form data search unit 7 (S 7: 1 case), the coincidence α made by the coincidence calculation unit 8 for the form is determined based on a predetermined criterion. If the value is equal to or greater than the value (S8: Yes), the content of the retrieved form is determined to be positive. In this case, the character replacement unit 10 replaces the data that could not be recognized by the character recognition unit 4 based on the form data searched by the form data search unit 7 (S9). A replacement flag is set for the replaced data.
[0064]
Further, in the character estimation unit 9, when a plurality of forms are searched by the form data search unit 7 (S7: 2 or more), the form of the form in which the degree of coincidence α made by the coincidence degree calculating unit 8 shows the maximum value. The content is positive (S18). In this case, the character replacement unit 10 replaces the data not recognized by the character recognition unit 4 based on the form data searched by the form data search unit 7 (S9). A replacement flag is set for the replaced data.
[0065]
The post-replacement character recognition data in which the data is replaced in this way is output from the character replacement unit 10 to the post-replacement character recognition data storage DB 11 and stored therein, and is also output to the display unit 12 (S10). .
[0066]
Even if the form is not searched by the form data search unit 7 (S7: 0) or there is a form searched by the form data search unit 7, it is calculated by the matching degree calculation unit 8. If the coincidence α is less than the predetermined reference value (S8: No), the data is not replaced. As described above, even when the data is not replaced, the character recognition data output from the character estimation unit 9 is output from the character replacement unit 10 to the display unit 12 (S11).
[0067]
The character recognition data output to the display unit 12 in this way is displayed on the screen together with the image data of this form stored in the image image data storage DB 3 (S12). In this case, the data replaced by the character replacement unit 10 is displayed using a different display color, a different character size, or a different modification method by recognizing that the replacement flag is “1”. Thereby, it is identified that the data is replaced. Further, in order to clarify the accuracy of the replaced data, it is also possible to display the colors separately according to the value of the matching degree α.
[0068]
In this way, the character recognition data and the image data corresponding to the form are displayed in parallel, so that the contents of both data are compared by the operator and whether the data is correctly recognized by the character recognition unit 4 or not. Alternatively, it is confirmed whether or not the data has been correctly replaced by the character replacement unit 10 (S13).
[0069]
As a result of the comparison between the two data, if the operator determines that the character recognition data includes erroneously recognized data, erroneously replaced data, or data that has not been character-recognized. If it is present (S14: Yes), the operator inputs change data from the change input unit 13 and the contents of the character recognition data are changed (S15).
[0070]
When data that is not character-recognized is not included (S14: No), or when change data is input from the operator or data that has not been recognized to the change input unit 13 in this way, Based on the input data, the contents of the character recognition data are updated and confirmed (S16).
[0071]
Character recognition data for which all data has been recognized in this way is output from the change input unit 13 to the form data storage DB 14 and stored as a form that has been subjected to character recognition processing in the past and data corresponding to each item of the form. (S17). The stored form and the data corresponding to each item of the form are used for the search performed in step S5.
[0072]
The character recognition processing system 1 described in the embodiment of the present invention can be realized by causing a computer to read a program stored in a storage medium.
[0073]
Examples of the storage medium include a magnetic disk, a floppy (registered trademark) disk, a hard disk, an optical disk (CD-ROM, DVD, etc.), a magneto-optical disk (MO, etc.), a semiconductor memory, etc. As long as it is a readable storage medium, the storage format may be any form.
[0074]
Further, an OS (operating system) operating on the computer based on an instruction of a program installed in the computer from the storage medium, MW (middleware) such as database management software, network software, or the like is an embodiment of the present invention. You may perform a part of each process for implement | achieving.
[0075]
Furthermore, the storage medium in the embodiment of the present invention is not limited to a medium independent of a computer, but also includes a storage medium in which a program transmitted via a LAN, the Internet, or the like is downloaded and stored or temporarily stored.
[0076]
Further, the number of storage media is not limited to one, and the case where the processing in the present embodiment is executed from a plurality of media is also included in the storage media in the present invention, and the media configuration may be any configuration.
[0077]
The computer according to the embodiment of the present invention executes each process according to the embodiment based on a program stored in a storage medium, and includes a single device such as a personal computer and a plurality of devices. Any configuration such as a network-connected system may be used.
[0078]
The computer in the embodiment is not limited to a personal computer, but includes an arithmetic processing device, a microcomputer, and the like included in an information processing device, and a device or device that can realize the functions of the embodiment by a program. Collectively.
[0079]
As described above, in the character recognition processing system according to the embodiment of the present invention, when there is unrecognized data as a result of the character recognition processing performed by the character recognition unit 4 due to the above-described operation, it is recognized. The search key can be arbitrarily selected from the items corresponding to the data. In this way, since the search key can be arbitrarily selected, it is possible to flexibly search a form with a high data matching rate from the previously processed forms, and the previously processed forms And data corresponding to each item can be used effectively.
[0080]
Based on the selected search key, a form having a high data matching rate can be searched from previously processed forms. Since the retrieved form is a form having a high data matching rate, the probability that the data replaced using this data is correct data is extremely high. Therefore, it is possible to prevent erroneous replacement and perform replacement with higher accuracy.
[0081]
Furthermore, by displaying the character recognized or replaced data together with the image data in which the data has been entered, it is confirmed that the character has been correctly recognized or correctly replaced and is necessary. In some cases, the contents can be modified.
[0082]
As described above, by using the character recognition processing system according to the embodiment of the present invention, even if there is a character that is not recognized from the image data of the form, the form processed in the past and its items By using corresponding data, it becomes possible to recognize unrecognized characters by estimating them with higher accuracy and replacing them with the estimated data. As a result, the operator's workload can be reduced.
[0083]
As mentioned above, although preferred embodiment of this invention was described referring an accompanying drawing, this invention is not limited to this structure. Within the scope of the invented technical idea of the scope of claims, a person skilled in the art can conceive of various changes and modifications, and the technical scope of the present invention also relates to these changes and modifications. It is understood that it belongs to.
[0084]
【The invention's effect】
As described above, according to the present invention, it is possible to arbitrarily set an item as a search key, and to effectively estimate data that has not been recognized in the character recognition process by using data accumulated in the past. Therefore, it is possible to realize a character recognition processing system and program capable of increasing the character recognition rate and reducing the burden on the operator.
[Brief description of the drawings]
FIG. 1 is a configuration diagram showing an example of the overall configuration of a character recognition processing system according to an embodiment of the present invention.
FIG. 2 is a diagram showing an example of a form subjected to character recognition processing
FIG. 3 is a diagram showing an example of a form retrieved from the form data storage DB and its data.
FIG. 4 is a diagram illustrating an example of a state in which a replacement flag is set in replaced data.
FIG. 5 is a flowchart showing the operation of the character recognition processing system according to the embodiment of the present invention.
FIG. 6 is a diagram illustrating an example of a form that has been subjected to character recognition processing;
FIG. 7 is a diagram showing an example of a form including unrecognizable data in data designated as a search key
[Explanation of symbols]
1 ... Character recognition processing system
2 ... Image data converter
3 ... Image image data storage database
4 ... Character recognition part
5. Character recognition data storage database
6 ... Search key selection part
7 ... Form data search part
8: Matching degree calculation unit
9 ... Character estimation part
10 ... Character replacement part
11 ... Character recognition data storage database after replacement
12 ... Display section
13 ... Change input section
14 ... Form data storage database

Claims (4)

Character recognition means for performing character recognition processing on each data of a form consisting of a plurality of items based on the image image of the form;
A form data storage means for storing a form whose character has been recognized in the past together with data corresponding to each item of the form;
If there is data that has not been character-recognized among the data that has been subjected to character recognition processing by the character recognition means, the form data storage means is based on the item of data that has been character-recognized by the character recognition means. A selection means for selecting the item to be a search key for searching for the form and the data of each item stored in the item from among items corresponding to the character-recognized data;
A search unit that searches the form data storage unit for a form including data corresponding to the item of the search key selected by the selection unit, and data corresponding to each item;
Estimation means for estimating the data that has not been recognized based on the form searched by the search means and the data corresponding to each item;
Replacing means for replacing the data not recognized by the character with the data estimated by the estimating means ;
When there are a plurality of forms searched by the search means, the degree of coincidence between the data corresponding to each item of the searched form and the data corresponding to each item of the form subjected to character recognition processing by the character recognition means And calculating means for calculating for each searched form,
The character recognition is characterized in that the estimation means estimates the data that has not been recognized based on data corresponding to each item of the form having the maximum degree of coincidence calculated by the calculation means. Processing system. The character recognition processing system according to claim 1 ,
Display means for displaying each data of the form together with an image image of the form by displaying the data replaced by the replacing means by using different display colors, different character sizes, or different modification methods;
A character recognition processing system, comprising: a change input unit that receives a change input of the replaced data displayed by the display unit. Character recognition processing is performed based on the data entered in the form corresponding to each item, and the data that has not been character-recognized is based on the form that has undergone character recognition processing in the past and the data corresponding to each item. A computer readable program that increases the character recognition rate by estimating its contents,
Character recognition means for performing character recognition processing on each data of a form composed of a plurality of items based on the image image of the form,
A form data storage means for storing a form whose character has been recognized in the past together with data corresponding to each item of the form;
If there is data that has not been character-recognized among the data that has been subjected to character recognition processing by the character recognition means, the form data storage means is based on the item of data that has been character-recognized by the character recognition means. A selection means for selecting the item to be a search key for searching the form and data of each item stored in the item corresponding to the character-recognized data;
Search means for searching a form including data corresponding to the item of the search key selected by the selection means and data corresponding to each item from the form data storage means;
Estimating means for estimating the data that has not been recognized based on the form searched by the searching means and the data corresponding to each item;
Replacing means for replacing the data not recognized by the character with the data estimated by the estimating means ;
When there are a plurality of forms searched by the search means, the degree of coincidence between the data corresponding to each item of the searched form and the data corresponding to each item of the form subjected to character recognition processing by the character recognition means Computing means for computing each retrieved form,
The estimation means includes means for estimating the data that has not been character-recognized based on data corresponding to each item of the form having the maximum degree of coincidence calculated by the calculation means. A program characterized by being realized. In the program according to claim 3,
Display means for displaying each data of the form together with an image image of the form by displaying the data replaced by the replacing means using different display colors, different character sizes, or different modification methods;
A program for causing a computer to realize change input means for receiving change input of the replaced data displayed by the display means.

Images