JP2017156922A - Document recognition control system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize a document recognition control system that converts information on wrong recognition parts to data upon reading documents to disclose the information thereof as the data, and points the wrong recognition parts out to operators.SOLUTION: A document recognition control system (1) comprises: document reading means (2); control means (3) that inputs a reading signal to convert information on a document (S) to data; and display means (4) that displays the information on the document (S). The control means (3) comprises: judgement means (6) that judges a content of the information on the document (S) thanks to the reading signal; recognition part-purpose first storage means (7) that, upon judging the content of the information on the document (S), converts information on recognition parts to data, and stores the information thereon as the data; wrong recognition part-purpose second storage means (8) that, upon judging the content of the information on the document (S), converts information on a wrong recognition part to data, and stores the information thereon as the data; and wrong recognition part point-out means (9) that points the information on the wrong recognition part out to display means (4) to display the information.SELECTED DRAWING: Figure 1

Description

  BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document recognition control system, and more particularly to a document recognition control system that discloses information on misrecognized portions, which are portions that could not be accurately recognized when optically reading a document, as data.

  At present, optical character recognition devices (also referred to as “OCR devices”) and optical mark reading devices (also referred to as “OMR devices”) are used, and information about documents including printed documents and handwritten documents is used by these devices. Is converted into data.

  And in patent document 1 mentioned later, the technique which correct | amends more correctly the misrecognition of the OCR apparatus by an entry character overlapping an entry frame is disclosed.

Japanese Patent Laying-Open No. 2015-97067

By the way, in the conventional OCR apparatus, when optically reading a document composed of a printed document or a handwritten document and converting it into data, a recognition portion that is a portion that has been successfully read and a portion that cannot be read are erroneously recognized. The part was disclosed uniformly.
For this reason, in the present situation, it is necessary to accurately create a document composed of a printed document or a handwritten document, so that more time and labor are consumed than necessary to create the document, and the usability of the OCR device is increased. However, it is disadvantageous in that it is disadvantageous for practical use.
In addition, it is necessary to develop a technique for improving the recognition accuracy at the time of optical reading processing in the OCR apparatus, and a technique for correcting the erroneous recognition of the OCR apparatus more accurately as in Patent Document 1 described above. There was a need to develop.
As a result, with the improvement of the recognition accuracy of the OCR device and the development of technology such as correcting the erroneous recognition of the OCR device more accurately, the OCR device becomes expensive and economically disadvantageous, There is an inconvenience that it cannot penetrate the market widely and its versatility is low.

  SUMMARY OF THE INVENTION An object of the present invention is to provide a document recognition control system that discloses information on misrecognized portions, which are portions that cannot be read when reading a document composed of a printed document or a handwritten document, as data, and points the misrecognized portion to the operator. It is to be realized.

  Therefore, in order to eliminate the inconvenience described above, the present invention provides a document recognition control system for optically reading a document consisting of a printed document or a handwritten document and controlling the read information as data. Means, a control means for inputting the reading signal from the document reading means to convert the information of the document into data, and a display means for displaying the information of the document by a control signal output from the control means, The control means determines the information content of the document based on a reading signal from the document reading means, and converts the information of the recognized portion into data when the determination means determines the content of the document information. The first storage means for the recognition location to be stored and the content of the information of the document is judged by the judgment means. A second storage means for misrecognized locations that stores the information of the location as data, and an error that indicates and discloses the information of the misrecognized locations stored in the second storage means for the misrecognized locations to the display means. A recognition location indicating means.

According to this invention, when optically reading a document consisting of a printed document or a handwritten document, the document recognition control system converts the information on the misrecognized portion, which is a portion that cannot be read, into data and indicates it to the operator. The operator can clearly grasp the misrecognized portion.
For this reason, misrecognized parts are pointed out after the reading process, and the misrecognized parts can be easily grasped and corrected, which reduces time and labor compared to the case of accurately creating documents. be able to.
Even if an OCR device that has not been developed as in the prior art is used, it is possible to use an inexpensive OCR device by easily correcting a misrecognized portion after the reading process.

FIG. 1 is a process diagram of a document recognition control system. Example 1 FIG. 2 is a schematic block diagram of a document recognition control system. Example 1 3A and 3B show an example of a document by the document recognition control system. FIG. 3A is a schematic diagram of the document before reading, and FIG. 3B is a schematic diagram of the document disclosed on the display means after reading. Example 1 FIG. 4 is a schematic block diagram of the document recognition control system. (Example 2) FIG. 5 shows an example of a document by the document recognition control system, (a) is a schematic diagram of a document which is a stylized print document before reading, and (b) is a schematic diagram of a document to which handwriting is added before reading. (C) is the schematic of the document disclosed by a display means after reading. (Example 2)

  Embodiments of the present invention will be described below in detail with reference to the drawings.

1 to 3 show Embodiment 1 of the present invention.
In FIG. 2, reference numeral 1 denotes a document recognition control system.
The document recognition control system 1 optically reads a document S, which is a printed document or a handwritten document, and performs control to convert the read information into data.

At this time, the document recognition control system 1 includes a document reading unit 2 that reads the document S, a control unit 3 that inputs a read signal from the document reading unit 2 and converts the information of the document S into data, Display means 4 for displaying the information of the document S in accordance with a control signal output from the control means 3.
The document reading means 2 uses a document reading function of an OCR device or OMR device, which is a known technique, and detailed description thereof is omitted.
The control means 3 stores a program for operating the document reading means 2 and is composed of a personal computer.
The display means 4 comprises a display.
The document recognition control system 1 has a character input function for correcting and joining the displayed characters when the information of the document S is displayed on the display means 4.
The document recognition control system 1 includes a printing unit 5 including a printer for printing the information of the document S displayed on the display unit 4.
Further, the control unit 3 determines the content of the information on the document S based on the read signal from the document reading unit 2, and the determination unit 6 determines the content of the information on the document S. First recognition means 7 for the recognition location that stores the information of the recognition location as data, and when the content of the information of the document S is determined by the determination device 6, the information of the erroneous recognition location is converted into data and stored. The second storage means 8 for the erroneously recognized location and the erroneously recognized location indicating means 9 for indicating the information of the erroneously recognized location stored in the second storage means 8 for the erroneously recognized location to the display means 4 and disclosing it. And a configuration comprising:
More specifically, the determination unit 6 of the control unit 3 stores basic determination character information in advance, and character data that is the content of the information on the document S is read by a read signal from the document reading unit 2. The character information for judgment is referred to, and the recognition state of the information is judged.
In addition, the first storage means 7 for the recognition location of the control means 3 can accurately read the document reading means 2 when the determination means 6 determines the character data that is the content of the information on the document S. The information of the recognized part, that is, the recognized part is converted into data and stored.
The second storage means 8 for the erroneously recognized portion of the control means 3 becomes incomplete in the reading operation of the document reading means 2 when the determining means 6 determines the character data that is the content of the information on the document S. The information of the part that could not be recognized, that is, the misrecognized part is stored as data.
Then, the misrecognized location indication means 9 of the control means 3 includes the information on the recognized location stored in the first storage means 7 for the recognized location and the second storage means 8 for the erroneously recognized location. When the information on the misrecognized portion stored in the above is disclosed, the information on the misrecognized portion in the second storage means 8 for the misrecognized portion is pointed out and disclosed.
In addition, as a measure of “pointing out and disclosing information on misrecognized portions” on the display means 4, as shown in FIG. 3B, character data that is the content of information on misrecognized portions is disclosed in bold characters. Measures to color or colorize.

In addition, when the control unit 3 determines the character data that is the content of the information of the document S by the determination unit 6, the misrecognition point indication unit increases as the misrecognition degree of the erroneous recognition unit increases. 9 includes a misrecognition degree adjusting unit 10 that enhances the display state of misrecognition information disclosed in the display unit 4 according to 9.
That is, when the judging means 6 judges the character data which is the content of the information of the document S having the already completed document format (printing recording medium) as shown in FIG. The degree of misrecognition is displayed at a location, that is, ranking is performed.
For example, ranking is performed in three stages as follows.
(1st rank)
: Character data that has the most misrecognized state and cannot be recognized at all.
(2nd rank)
: In character data, misrecognition occurs, but the corresponding character can be disclosed from the character information for determination with a probability of about 50 to 70%.
(3rd rank)
: In character data, misrecognition occurs, but the corresponding character can be disclosed from the character information for determination with a probability of about 30%.
And according to each rank, a bold character or coloring is performed, and each rank can be discriminated visually.

That is, the document recognition control system 1 includes the following steps as shown in FIG.
A: <Document reading process>
B: <Content judging process of control means>
C: <First storage step of control means>
D: <Second storage step of control means>
E: <Display process>
F: <Correction process>

  Next, processing steps of the document recognition control system 1 will be described with reference to FIG.

A: <Document reading process>
In this <document reading step>, the document reading means 2 optically reads the document S composed of a printed document or a handwritten document.
B: <Content judging process of control means>
In this <content determination process of the control means>, the determination means 6 of the control means 3 determines the document S based on the determination character information stored in advance in the determination means 6 and the read signal from the document reading means 2. The character data that is the content of the information is inquired to determine the information recognition state.
C: <First storage step of control means>
In this <first storage step of the control means>, when the judgment means 6 judges the character data that is the content of the information on the document S, the reading work of the document reading means 2 is accurately performed and recognized. The information on the completed location, that is, the recognized location is converted into data and stored in the first storage means 7 for the recognized location of the control means 3.
D: <Second storage step of control means>
In this <second storage step of the control means>, when the character data that is the content of the information on the document S is determined by the determination means 6, the reading operation of the document reading means 2 is incomplete and cannot be recognized. The information of the detected location, that is, the erroneously recognized location is converted into data and stored in the second storage means 8 for the erroneously recognized location of the control means 3.
E: <Display process>
In this <display step>, information on the recognition location stored in the first storage means 7 for the recognition location on the display means 4 and the erroneous recognition location stored in the second storage means 8 for the erroneous recognition location. Information.
At this time, the misrecognized location indication means 9 of the control means 3 points out and discloses the information of the misrecognized location in the second storage means 8 for the misrecognized location.
As shown in FIG. 3A, when the character data which is the content of the information of the document S having the already completed document format (printing recording medium) is “C”, the character data “ When the second storage means 8 for erroneous recognition locations of the control means 3 determines that “C” is an erroneous recognition location, the character data closest to the character data “C” is disclosed to the display means 4 As shown in FIG. 3B, the document S, for example, “G” is disclosed in bold letters in consideration of ranking.
At this time, in this <display step>, in addition to a measure for disclosing character data in bold characters according to the ranking, a measure for coloring may be considered.
F: <Correction process>
In this <correction step>, the erroneous recognition location disclosed in the display means 4 is corrected by the character input function of the document recognition control system 1.

Thus, according to the first embodiment, when the document S, which is a printed document or a handwritten document, is optically read, the document recognition control system 1 stores information on a misrecognized portion that cannot be read as data. By making it to the operator and making it point out, the operator can clearly grasp the erroneously recognized part.
For this reason, a misrecognized portion is pointed out after the reading process, and the misrecognized portion can be easily grasped and corrected. Compared with the case where the document S is accurately created, time and labor are saved. It can be reduced.
Even if an OCR device that has not been developed as in the prior art is used, it is possible to use an inexpensive OCR device by easily correcting a misrecognized portion after the reading process.

Further, by providing the control means 3 with the misrecognition degree adjusting means 10 for increasing the display state of misrecognition information, the misrecognition degree adjusting means 10 according to the misrecognition degree of the misrecognized portion of the document S, for example. By means of this, it is possible to rank misrecognized portions in the character data that is the content of the information of the document S when it is disclosed to the display means 4, and it can be visually discriminated by each rank by bold characters or coloring. Become.
For this reason, ranking of misrecognized portions of the document S can contribute to reduction of labor and time in the content confirmation work of the information of the document S.

4 and 5 show Embodiment 2 of the present invention.
In the second embodiment, portions having the same functions as those in the first embodiment will be described with the same reference numerals.

  The feature of this second embodiment is that the control means 12 of the document recognition control system 11 is provided with a print document data storage means 13, and the print data stored in advance in the print document data storage means 13 is erroneously recognized. The point is that it is configured to be displayed for reference.

That is, as shown in FIG. 4, the document recognition control system 11 includes the document reading means 2, the control means 12, and the display means 4 as in the first embodiment.
The document recognition control system 11 includes the printing unit 5.
Further, the control means 12 includes the determination means 6, a first storage means 7 for the recognition location, a second storage means 8 for the erroneous recognition location, and the erroneous recognition location indication means 9. ing.
At this time, as shown in FIG. 4, the control means 12 includes the print document data storage means 13 for preliminarily storing information on the document SS when the stylized print document is used as the document SS.
When the control means 12 determines the content of the information of the document SS by the determination means 6, the control means 12 inquires the print data stored in the print document data storage means 13 in advance for the misrecognized location, and The print data stored in advance in the document data storage means 13 is configured to be displayed as a reference at a misrecognized location.

In other words, before the reading, as shown in FIG. 5A, the document SS composed of the standardized print document is stored in the print document data storage means 13 in advance.
Then, as shown in FIG. 5 (b), handwriting is performed on the necessary part of the document SS.
After that, when the document SS read by the document reading means 2 is disclosed to the display means 4, as shown in FIG. 5 (c), a necessary portion which is a handwritten portion becomes an erroneously recognized portion. Is inquired of the print data stored in advance in the print document data storage means 13, and the print data stored in advance in the print document data storage means 13 is displayed as a reference at a misrecognized location.
That is, as shown in FIG. 5B, for example, the number “4” in the first line of the document SS, the number “6” in the second line, and the number “3” in the third line. When the number “2” in the fourth line and the number “8” in the fifth line are erased by handwriting with a writing tool, as shown in FIG. For reference, “4”, “6”, “3”, “2”, and “8” are displayed in bold.

Thus, according to the second embodiment, when optically reading the document SS made up of a stylized print document, a necessary portion that is a handwritten portion becomes a misrecognized portion, and the document recognition control is performed. The system 11 makes it possible for the operator to clearly grasp the misrecognized part by displaying the information of the misrecognized part, which is a part that cannot be read, as a reference display and pointing out to the operator.
For this reason, a necessary location is displayed as a misrecognized location after the reading process, the misrecognized location can be easily grasped, and the time and labor when creating the document SS can be reduced. .
In addition, if the misrecognized portion is displayed in black or colored bold characters, it can be converted into data as an important portion of the document SS made up of a stylized printed document, for example, various reports Data storage of such a large amount of documents SS can be easily performed.
Furthermore, even if an OCR device that has not been developed as in the prior art is used, it is possible to use an inexpensive OCR device by easily grasping a misrecognized portion after the reading process.

In addition, the document recognition control system 11 according to the second embodiment is added with the misrecognition degree adjusting means 14 for enhancing the display state of the misrecognition information to the control means 12 described in the first embodiment. It is also possible to have a configuration in which erroneous recognition locations in the character data that is the contents of the information of the document SS are ranked when a predetermined location of the SS cannot be intentionally recognized by handwriting.
In other words, as shown in FIG. 5 (a), when the document SS is configured with another handwritten document, for example, a stylized print document, and a document format (handwritten recording medium) that has already been completed, When the character data which is the content of the information of the document SS is determined by the determination means 6, the degree of erroneous recognition, that is, ranking is performed on the erroneous recognition portion.
At this time, as shown in FIG. 5 (a), when a completed document format (recording medium for handwriting) is created, the characters described in this document format are written as shown in FIG. 5 (b). Erase by hand with a tool, and process this erased location as a misrecognized location.
Then, as in the first embodiment, ranking is performed in three stages as follows.
(1st rank)
: Character data that has the most misrecognized state and cannot be recognized at all.
(2nd rank)
: In character data, misrecognition occurs, but the corresponding character can be disclosed from the character information for determination with a probability of about 50 to 70%.
(3rd rank)
: In character data, misrecognition occurs, but the corresponding character can be disclosed from the character information for determination with a probability of about 30%.
And according to each rank, a bold character or coloring is performed, and each rank can be discriminated visually.
Then, by providing the control means 3 with the misrecognition degree adjusting means 14 for increasing the display state of misrecognition information, for example, a predetermined part of the document SS having a document format (recording medium for handwriting) that has already been completed. Since the misrecognition degree of the misrecognized portion increases according to the deliberate degree, the misrecognition degree adjusting unit 10 causes the misrecognition degree to be disclosed to the display unit 4 It is possible to rank misrecognized portions in the character data that is the content of the information of the document SS, and it is possible to visually discriminate by each rank by bold characters or coloring.
For this reason, it is possible to rank according to the priority order for confirming the document SS, and it is also possible to rank when confirming the content of the information of the document SS. It can contribute to the reduction.

  The present invention is not limited to the first and second embodiments described above, and various application modifications can be made.

For example, in the first embodiment of the present invention, the display means of the document recognition control system is character data that is the closest to the character data “C”, for example, “G” in bold characters in consideration of ranking. Although the document has been disclosed, a special configuration in which a pickup function is added to the document recognition control system is also possible.
That is, in the first embodiment of the present invention, as is apparent from FIGS. 3A and 3B, when documents are disclosed on the display means, both are displayed in the same size. However, this pick-up function also changes the size of the displayed character according to the degree of misrecognition of the misrecognized portion of the document.
Then, the misrecognized portions of the document are displayed in a size, thickness, or color according to the degree of misrecognition, so the content of the information on the document is confirmed. Ranking can be made, and it can contribute to the reduction of labor and time in the work of checking the contents of document information.
At this time, in the display operation by the pickup function described above, it is always displayed only when the cursor points out a measure for changing the size of the displayed character according to the degree of misrecognition of the misrecognized portion of the document. A function for enlarging the pickup display, which includes measures to change the size of the characters to be displayed, can be considered.

DESCRIPTION OF SYMBOLS 1 Document recognition control system 2 Document reading means 3 Control means 4 Display means 5 Printing means 6 Judgment means 7 1st memory means for recognition location 8 Second storage means for erroneous recognition location 9 Error recognition location indication means S Document

Claims (3)

  In a document recognition control system for optically reading a document consisting of a printed document or a handwritten document and controlling the read information to be converted into data, a document reading means for reading the document and a reading signal from the document reading means are input. Control means for converting the information of the document into data, and display means for displaying the information of the document by a control signal output from the control means. The control means is based on a reading signal from the document reading means. Determining means for determining the content of the information of the document, and when the content of the information of the document is determined by the determining means, the first storage means for the recognition location that stores the information of the recognition location as data; When the content of the information of the document is judged by the judging means, the information for the misrecognized part is stored as data of the misrecognized part as data. Document recognition comprising: storage means; and misrecognition location indicating means for indicating and disclosing information on misrecognition locations stored in the second storage means for this erroneous recognition location to the display means Control system.   When the control means determines the content of the document information by the determination means, the misrecognition disclosed to the display means by the misrecognition location indication means as the misrecognition location misrecognition degree increases. The document recognition control system according to claim 1, further comprising a misrecognition degree adjusting unit for increasing a display state of information.   The control means includes a print document data storage means for preliminarily storing information on the document when the stylized print document is used as the document, and when the content of the document information is determined by the determination means. The print data stored in advance in the print document data storage means is referred to the print data stored in the print document data storage means, and the print data stored in the print document data storage means is displayed as a reference in the error recognition position. The document recognition control system according to 1.

Images