JP2000187704A - Character recognition device, its method and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make it possible to make a user more accurately know characters whose similarity is uncertain and to improve the usability of the character recognition device. SOLUTION: When an image is inputted in a step S31, a character is segmented from the inputted image in a step S32. In a step S33, character recognition processing is applied to each character segmented in the step S32 to obtain respective similarity. In a step S34, the similarity obtained by the character recognition processing in the step S33 is corrected based on the size (segmenting size) of each character segmented in the step S32. In the case of displaying a recognized result in a step S35, the recognized result is provided to a user by changing the colors of characters e.g. so as to discriminate a recognized result whose similarity is lower than a prescribed threshold.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character recognition apparatus and method for recognizing characters based on a pattern existing in input image data.

[0002]

2. Description of the Related Art In general, a character recognition device divides a character image into several parts, extracts a directional component of the character for each area, and compares the directional component with a directional component of a character to be recognized which is stored in advance to determine the similarity. It calculated and output the recognition result.

[0003]

Generally, in character recognition, a character pattern needs to have a certain size in order to accurately calculate similarity. That is, in the character recognition process, there are cases where the character is large and the similarity can be accurately output, and cases where the character is small and crushed and the similarity cannot be output very accurately. In particular, superscripts, punctuation marks, and the like often have small characters, tend to be crushed, and cannot provide a similarity very accurately.

In general, in character recognition processing, there is no distinction between a case where a character is able to give a large degree of similarity accurately and a case where the similarity is not accurate. It was difficult to inform and to give special treatment to uncertain characters.

SUMMARY OF THE INVENTION The present invention has been made in view of the above conventional example, and has as its object to provide a character recognition apparatus and a method thereof that can inform a user of a character whose similarity is uncertain.

It is another object of the present invention to efficiently perform processing for reducing erroneous recognition and improve recognition accuracy.

[0007]

In order to achieve the above object, a character recognition device according to the present invention has, for example, the following configuration. That is, a character recognition device that performs character recognition based on an input image, a cutout unit that cuts out a character from the input image, and performs a character recognition process on the character cutout by the cutout unit. A recognition unit that obtains each similarity; and a correction unit that corrects the similarity obtained by the recognition unit based on the size of the character extracted by the extraction unit.

A character recognition device according to the present invention for achieving the above object has, for example, the following configuration. That is, a character recognition device that performs character recognition based on an input image, a cutout unit that cuts out a character from the input image, and performs a character recognition process on the character cutout by the cutout unit. A recognition unit that obtains each similarity; and a correction unit that corrects the similarity obtained by the recognition unit based on the position of the character extracted by the extraction unit.

Preferably, the character recognition apparatus further includes a correction unit that corrects a recognition result having a low similarity after correction by the correction unit based on the recognition results around the correction result.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Some preferred embodiments of the present invention will be described below in detail with reference to the accompanying drawings.

<Description of Character Recognition Apparatus (FIG. 1)> FIG. 1 is a block diagram showing a schematic configuration of a character recognition apparatus according to an embodiment of the present invention. In FIG. 1, reference numeral 1 denotes the entire character recognition device of the present embodiment. Reference numeral 2 denotes an input unit for inputting an image, for example, inputting an image stored in a hard disk or the like. Alternatively, the input unit 2 may be a scanner that optically reads a document image. Reference numeral 3 denotes a central processing unit (hereinafter, referred to as a CPU) for arithmetic processing, which controls the entire character recognition device 1.

Reference numeral 4 denotes a ROM, which stores control programs and various data shown in the flowcharts of FIG. Also, a recognition dictionary 9 used in character recognition is stored. Reference numeral 5 denotes a memory (RAM), which is used as a work area of the CPU 3, temporarily stores image data input from the input unit 2, and stores the position and size of each character extracted by the character cutting unit 7. And a recognition unit 9
There is also provided an area for storing the candidate character and similarity for each character calculated in.

The CPU 3 executes various control programs stored in the ROM 4 to thereby execute the similarity correction section 6,
The character cutting unit 7, the recognition unit 8, and the display unit 10 are realized. The similarity correction unit 6 corrects the similarity using the character size information. The character cutting unit 7 extracts the characters of the text in the image stored in the memory 5 and stores the positions and sizes of the characters in the memory 5. The recognition unit 8 performs character recognition using the recognition dictionary 9. The display unit 10 controls display on the display 14 based on the input image and the character recognition result.

A system bus 11 has a CPU
3 including a data bus, an address bus and a control signal bus. Reference numeral 12 denotes an interface unit that controls an interface with an external output device, for example, a printer 13 or a display 14.

<First Embodiment> FIG. 2 is a view showing an example of a text window showing the result of character recognition processing by the character recognition device 1 of the first embodiment. In FIG.
Reference numeral 21 denotes one character of the normal recognition result, which is displayed in black. Reference numeral 22 denotes an uncertain recognition result, that is, one character of a recognition result having a high possibility of being erroneous, and is displayed in red. Hereinafter, a control procedure for realizing such display will be described.

FIG. 3 is a flowchart for explaining the procedure of the character recognition process according to the first embodiment. As described above, the control program for realizing this processing is the ROM
4 is executed by the CPU 3.

First, in step S31, an image is input to the input unit 2.
The data is read from a hard disk or a scanner, input, and stored in the memory 5.

Then, the process proceeds to a step S32, wherein the character cutting unit 7
Cuts out a character to be recognized. Any known method may be used as a method for cutting out characters.
For example, in the case of a horizontally written sentence, a horizontal histogram of black pixels of the document image is taken, an area where black pixels are continuously present is taken out as a row, and then a vertical histogram is taken for a row portion and black pixels are continuously taken. Extract existing area as character.

Next, the operation proceeds to step S33, in which the recognition unit 8 performs character recognition of the character extracted in step S32.

Then, the process proceeds to step S34, in which the similarity correction unit 6 corrects the similarity obtained in step S33 using the character size information obtained in step S32. Here, the correction of the similarity is performed, for example, as follows.

In the present embodiment, the correction rate is set according to the size of the character, for example, as follows, and the similarity is corrected. That is, For example, when a similarity of 8000, 32 dots vertically and 14 dots horizontally is obtained for the character 22 (comma “,”) in FIG. 2, the similarity of the character 22 (comma “,”) is 8000.
× 0.6 = 4800, which is corrected.

The above correction processing is performed on all the characters obtained in step S33, and the result is stored in the memory 5.

Next, the process proceeds to step S35, where the display unit 10 displays the recognition result as shown in FIG.

Here, for example, a character having a similarity of less than 7000 is displayed as a red character as an uncertain recognition result, that is, a recognition result having a high possibility of being erroneous.
As a result, the character 22 is displayed in red as shown in FIG. On the other hand, characters having a similarity of 7000 or more are displayed as black characters like character 21 in FIG.

The description of the process of outputting an image or the like from the printer 13 is omitted.

As described above, according to the first embodiment,
By correcting the similarity calculated based on the size of the recognized character, it is possible to accurately inform the user of the uncertain character. Therefore, it is easy to correct misrecognition,
A character recognition device that is easy for the user to use can be realized.

<Second Embodiment> In the first embodiment, after the similarity is corrected in step S34 of FIG. 3, the result is displayed in step S35. In the second embodiment, an attempt is made to reduce erroneous recognition by correcting a recognition result whose similarity does not exceed the threshold value.

FIG. 4 is a flowchart for explaining the procedure of character recognition according to the second embodiment. A control program for realizing this processing is stored in the ROM 4,
It is executed by the CPU 3. In FIG. 4, the same processing as in the first embodiment (FIG. 3) is performed in steps S31 to S35. In the second embodiment, after the similarity correction processing in step S34, in step S41, individual processing is performed to reduce erroneous recognition of a recognition result with a low degree of recognition. Hereinafter, the process in step S41 will be described.

Similar to the first embodiment, similarity 7000
Characters (recognition results) less than
Individual processing is performed on these recognition results.

For example, the character 22 comma “,” in FIG. 2 whose similarity has been corrected in the first embodiment is subjected to individual processing because the similarity is 4,800.

An example of the individual processing will be described below. Note that the similarity is not changed and is not changed. Therefore, in the display processing in step S35, the character individual processing content is changed to "," if the end of the sentence is ".". If the end of the sentence is ".", It is changed to ",". Yes "." If there is "," in the middle of the sentence, change it to ".". "." If there is a "," in the middle of the sentence, change it to "." ... Individual processing according to the above example For example, in “Today is fine weather” before the individual processing, the similarity of “,” is less than a predetermined value, so the individual processing is executed and changed to “,”, and “Today is fine weather”. It becomes. "

In step S35, characters having a similarity of less than 7000 are displayed in red, as in the first embodiment.

As described above, according to the present embodiment,
The user can be notified of uncertain characters accurately, and individual processing for reducing erroneous recognition can be efficiently performed, thereby improving usability such as improving recognition accuracy.

<Third Embodiment> Next, a third embodiment will be described. FIG. 5 is a diagram illustrating a display state of a recognition result according to the third embodiment. In the first and second embodiments, the similarity is corrected based on the size of each character to be recognized (the size of the cut-out character).
In the embodiment, the similarity correction is performed based on the position of the character to be recognized.

FIG. 6 is a flowchart for explaining a character recognition procedure according to the third embodiment. Note that a control program for realizing this processing is stored in the ROM 4 and executed by the CPU 3. Step S3
1, S32, S33, and S35 are the first embodiment (FIG. 3)
Since the processing is the same as described above, the description is omitted here.

In step S61, the degree of similarity obtained in step S33 is corrected as follows using the character position information obtained in step S32.

Position information Correction rate Superscript 0.8 Subscript 0.8 Ruby 0.6 Tate-chu-yokoji 0.9 * Tate-chu-yoko is horizontal characters such as two characters included in vertical writing. These are characters written side by side.

For example, it is assumed that the similarity of the character 52 (superscript parentheses "(") shown in Fig. 5 is 8000 as a result of character recognition by the step S33 (recognition unit 8). Based on the cutout position, it is determined that the parenthesis is a parenthesis “(”, so the similarity is corrected to 8000 × 0.8 = 6400.

The above processing is performed on all the characters obtained in step S33, and the result is stored in the memory 5.

Next, the process proceeds to step S35, where the recognition result is displayed.

Here, for example, a character having a similarity of less than 7000 is displayed in red in the recognition result as an uncertain recognition result, that is, a recognition result having a high possibility of being erroneous. For example, the character 52 in FIG. 5 is displayed in red. On the other hand, characters having a similarity of 7000 or more are displayed as black characters. For example, in the recognition result of FIG.
1 is displayed in black letters.

The description of the process of outputting an image or the like from the printer 13 is omitted.

As described above, according to the third embodiment,
It is possible to inform the user of the uncertain character more accurately, and it is possible to realize a character recognition device that can be easily corrected for erroneous recognition and is easy for the user to use.

<Fourth Embodiment> Next, a fourth embodiment will be described. In the fourth embodiment, individual processing for reducing erroneous recognition is performed based on the similarity correction processing for character recognition based on the character position in step S61.

FIG. 7 is a flowchart for explaining the procedure of character recognition processing according to the fourth embodiment. Note that a control program for realizing this processing is stored in the ROM 4 and executed by the CPU 3.

In FIG. 7, after the similarity is corrected in step S61 according to the above-described third embodiment, in step S71, individual processing for correcting erroneous recognition for correcting the recognition result is performed. In step S71, for example,
Individual processing according to the recognition result is performed on characters having a similarity of less than 7000. The similarity is not changed.

For example, the superscript parenthesis "(" corrected for similarity in the third embodiment is subject to individual processing since the similarity is 6400. Some examples of individual processing are shown below.

Characters: Individual processing contents (: In the following sentence, if there is “]” with a higher similarity, “[”. If “｝” with a higher similarity is “｛”, If there is a higher “>”, change it to “<”. [: If the following sentence contains “)” with a higher similarity, “(” If ">" has a higher similarity to "", change it to "<". ｛: If ")" has a higher similarity in the following sentence, change it to ">" for a higher similarity. If there is, change it to “<”, and if there is a higher similarity “]”, change it to “[”. If there is a high “)”, change it to “(”. If there is a higher similarity “｝”, change it to “｛”.

By performing the above processing, for example, before the individual processing, a recognition result of “Today (2) is fine weather” is obtained, and the similarity of “(” is 6400 ≦ “]” and the similarity is 7
If it is 100, individual processing is executed. By this individual processing, "(" is changed to "[" and "Today [2]
Is sunny. ".

In step S35, characters having a similarity of less than 7000 are displayed in red, as in the third embodiment.

It should be noted that the processing described in the second embodiment can be applied as the individual processing. Further, the processing described in the fourth embodiment can be used in the second embodiment. Of course, the processes described in the second and fourth embodiments may coexist.

As described above, according to the present embodiment,
The user can be notified of an uncertain character accurately, and individual processing for reducing erroneous recognition can be efficiently performed, and the recognition accuracy can be increased. As a result, usability is improved.

Even if the present invention is applied to a system including a plurality of devices (for example, a host computer, an interface device, a reader, a printer, etc.), an apparatus (for example, a copier, a facsimile, etc.) comprising one device. Device).

It is another object of the present invention to provide a storage medium storing a program code of software for realizing the functions of the above-described embodiments to a system or an apparatus, and to provide a computer (or CPU) of the system or the apparatus.
And MPU) read and execute the program code stored in the storage medium.

In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiment, and the storage medium storing the program code constitutes the present invention.

As a storage medium for supplying the program code, for example, a floppy disk, hard disk, optical disk, magneto-optical disk, CD-ROM, CD
-R, a magnetic tape, a nonvolatile memory card, a ROM, or the like can be used.

When the computer executes the readout program code, not only the functions of the above-described embodiment are realized, but also the OS (Operating System) running on the computer based on the instruction of the program code. ) May perform some or all of the actual processing, and the processing may realize the functions of the above-described embodiments.

Further, after the program code read from the storage medium is written into a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, based on the instructions of the program code, It goes without saying that the CPU provided in the function expansion board or the function expansion unit performs part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

[0059]

As described above, according to the present invention,
It is possible to more accurately notify the user of a character whose similarity is uncertain. Further, according to the present invention, it is possible to efficiently perform processing for reducing erroneous recognition and improve recognition accuracy. Therefore, usability of the character recognition device is improved.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of a character recognition device according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating an example of a text window showing a result of a character recognition process performed by the character recognition device 1 according to the first embodiment.

FIG. 3 is a flowchart illustrating a procedure of a character recognition process according to the first embodiment.

FIG. 4 is a flowchart illustrating a procedure of character recognition according to a second embodiment.

FIG. 5 is a diagram illustrating a display state of a recognition result according to a third embodiment.

FIG. 6 is a flowchart illustrating a character recognition procedure according to a third embodiment.

FIG. 7 is a flowchart illustrating a procedure of a character recognition process according to a fourth embodiment.

[Explanation of symbols]

 Reference Signs List 1 character recognition device 2 input unit 3 CPU 4 ROM 5 memory 6 similarity correction unit 7 character cutting unit 8 recognition unit 9 recognition dictionary 10 display unit 11 system bus 12 interface unit 13 printer 14 display

Claims (22)

[Claims] 1. A character recognition device for performing character recognition based on an input image, comprising: a cutout unit that cuts out a character from the input image; and a character recognition process for the character cutout by the cutout unit. And a correction unit that corrects the similarity obtained by the recognition unit based on the size of the character extracted by the extraction unit. Character recognition device. 2. The character recognition device according to claim 1, further comprising a presentation unit that presents a character recognition result based on the similarity after the correction by the correction unit. 3. The character recognition device according to claim 2, wherein the presentation unit displays a recognition result having a similarity smaller than a predetermined value so that the user can identify the recognition result. 4. The character recognition device according to claim 1, wherein the correction unit corrects the character cut out by the extraction unit so as to reduce the degree of similarity with respect to the small-sized character. 5. The method according to claim 1, wherein the correcting unit is provided with a plurality of ranks for the size cut out by the cutting unit, and the ratio for changing the similarity is different for each rank. The character recognition device according to claim 1. 6. The character recognition apparatus according to claim 1, further comprising a correction unit that corrects a recognition result having a low degree of similarity after correction by the correction unit based on a recognition result around the correction result. 7. The character recognition device according to claim 6, wherein the correction unit performs correction based on consistency of punctuation marks. 8. The character recognition device according to claim 6, wherein the correction unit performs correction based on matching of parentheses. 9. A character recognition apparatus for performing character recognition based on an input image, comprising: a cutout unit that cuts out a character from the input image; and a character recognition process for the character cutout by the cutout unit. And a correction unit for correcting the similarity obtained by the recognition unit based on the position of the character extracted by the extraction unit. Recognition device. 10. The character recognition apparatus according to claim 9, further comprising a presentation unit that presents a character recognition result based on the similarity after the correction by the correction unit. 11. The character recognition apparatus according to claim 10, wherein the presentation unit displays a recognition result having a similarity smaller than a predetermined value so that the user can identify the recognition result. 12. The apparatus according to claim 1, wherein the correction unit corrects the position extracted by the extraction unit so as to reduce the degree of similarity with respect to a character at a specific position having a smaller character size than usual. Item 10. The character recognition device according to item 9. 13. The correction means, wherein the determined position includes at least one of a superscript, a subscript, ruby, and tate-chu-yoko, and sets a ratio for changing the similarity for each determined position. The character recognition device according to claim 9, wherein the character recognition device is associated with the character recognition device. 14. The character recognition apparatus according to claim 9, further comprising a correction unit that corrects a recognition result having a low similarity after correction by the correction unit based on a recognition result around the recognition result. 15. The character recognition device according to claim 14, wherein the correction unit performs correction based on consistency of punctuation marks. 16. The character recognition device according to claim 14, wherein the correction unit performs correction based on matching of parentheses. 17. A character recognition method for performing character recognition based on an input image, comprising: a cutout step of cutting out characters from the input image; and a character recognition process for the characters cut out by the cutout step. And a correction step of correcting the similarity obtained in the recognition step based on the size of the character cut out in the cut-out step. Character recognition method. 18. A character recognition method for performing character recognition based on an input image, comprising: a cutout step of cutting out characters from the input image; and a character recognition process for the characters cut out by the cutout step. And a correction step of correcting the similarity obtained in the recognition step based on the position of the character cut out in the cut-out step. Recognition method. 19. The apparatus according to claim 17, further comprising a correction step of correcting a recognition result having a low degree of similarity after the correction in the correction step based on the recognition results around the recognition result.
Or the character recognition method described in 18. 20. A storage medium for storing a control program for causing a computer to perform character recognition based on an input image, the control program comprising: a code for an extraction step for extracting characters from the input image. And performing a character recognition process on the character cut out in the cut-out step to obtain a similarity between each character, and a code obtained in the recognition step based on the size of the character cut out in the cut-out step. A correction process code for correcting the obtained similarity. 21. A storage medium for storing a control program for causing a computer to perform character recognition based on an input image, the control program comprising: a code for a cutout step of cutting out characters from the input image And performing a character recognition process on the character extracted by the extraction process to obtain a similarity between the code and a position of the character extracted by the extraction process. A correction process code for correcting the similarity. 22. The method according to claim 20, further comprising a code of a correction step of correcting a recognition result having a low similarity after the correction in the correction step based on a recognition result around the correction result. Storage medium.