JPH04106680A - Character reading and correcting device - Google Patents

Abstract

PURPOSE:To improve the processing capacity of the title device by displaying correct candidate characters from a candidate character storing memory by pressing a key and selecting a correct character category from the displayed correct candidate character categories. CONSTITUTION:Candidate character categories read by a character reading section (OCR) 1 are successively transferred to a read character storing buffer 2 and a recognition controlling section 3 successively transfers character categories corresponding to individual character images to a display controlling section 4 after reading out the character categories. The section 4 generates corresponding character patterns and transfers the patterns to a displaying section 5 for display. When an operator finds a wrong character on the section 5, the operator presses a candidate character selecting key 61 after moving a cursor to the position of the wrong character by using a keyboard 6. Then a code conversion circuit 7 transfers a candidate character selection instructing signal to the section 4. Accordingly, the section 4 sends another candidate character category corresponding to the same character image to the section 5 and the section 5 displays a character pattern corresponding to the candidate character category. When the character pattern is correct, the operator transfers a correction command signal to the section 3 by pressing a correction key 62. Therefore, correcting efficiency of this character reading and correcting device is improved.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a character reading/correcting device that corrects character categories read by a character reading device through operator intervention.

[Conventional technology]

Conventionally, devices that optically read printed or handwritten characters (hereinafter referred to as OCR) have been put into practical use for alphanumeric characters, hiragana, and katakana, and recently, devices that can read printed and handwritten kanji have also appeared. There is. As described above, when the range of characters to be read is expanded to include Chinese characters, it becomes difficult to obtain sufficient recognition accuracy due to the increase in the number of characters to be read and the complexity of character shapes. Therefore, it is necessary to correct the results read by OCR.

In the conventional correction method, a candidate character category consisting of one character is shown for one character image read by OCR, and the correct answer is selected from the character candidate category.

[Problem to be solved by the invention]

However, with this method, it is fundamentally impossible to present the candidate character categories obtained from the character reading device to the operator for images generated by contact or for character images obtained by combining multiple character images. However, the function as a character correction device was insufficient. Therefore, in order to correct the characters read by the character reading device, the operator must take steps such as inputting the character category again, which causes an overall decrease in the processing performance of the OCR system. .

SUMMARY OF THE INVENTION An object of the present invention is to provide a character reading/correcting device that eliminates the drawbacks of the prior art described above.

[Means for Solving the Problems] The character reading correction device of the present invention has a candidate character storage memory and corrects the character reading result through operator intervention, which reads characters from a plurality of character images. means for storing in the candidate character storage memory a character category string consisting of a plurality of character categories obtained using the device and a character category obtained using the character reading device from a character image obtained by combining the plurality of character images; , means for correcting a character category by an operator selecting either a candidate character category or a candidate character category string in the candidate character storage memory as a correct character category for an incorrect character category. shall be.

[Operation] In the present invention, a candidate character storage memory for storing candidate character categories obtained from a character reading device for a character image obtained by combining a plurality of character images is prepared and stored therein, and a predetermined key is pressed. This solves the above problem by displaying correct character candidates from the candidate character storage memory and providing a mechanism for selecting the correct character category from the displayed candidate character categories.
Improved the processing power of the R system.

〔Example〕

A specific embodiment of the present invention will be described below using a functional block diagram.

This character reading correction device is a character reading section (OCR).
) Read character storage buffer 2 that stores candidate character categories obtained from the character image read by 1.
, a recognition control section 3 , a display control section 4 , a display section 5 , a keyboard 6 , and a code conversion circuit 7 .

The read character storage buffer 2 contains candidate character categories obtained from 0CR1 for a character image that is a composite of multiple character images, and character categories obtained from 0CR1 for individual character images when character images are not composited. shall be stored in correspondence.

As the recognition control section 3, it is appropriate to use an ordinary microcomputer.

Two suitable keys on the keyboard 6 are assigned special codes in advance, and candidate character selection keys 61. This is a modification key 62. In this embodiment, a keyboard is used, but an input unit such as a tablet may also be used.

The code conversion circuit 7 uses candidate character selection keys 61 .
It is detected whether or not the correction key 62 has been pressed, and when a code indicating that the candidate character selection key 61 has been pressed is input, a candidate character selection command signal is output to the display control unit 4 to indicate that the correction key 62 has been pressed. When the code is input, a correction command signal is output to the recognition control section 3.

Next, the operation of this embodiment will be explained.

The candidate character categories read by 0CR1 are sequentially transferred to the read character storage buffer 2. The recognition control unit 3 sequentially reads character categories corresponding to individual character images from the read character storage buffer 2 and transfers them to the display control unit 4.

The display control section 4 generates a character pattern corresponding to the character category sent from the recognition control section 3, transfers it to the display section 5, and displays it. Note that when displaying the character categories read by the OCR 1, it is also possible to simultaneously display input character images corresponding to the reading results from the pattern memory built into the OCR 1.

When the operator checks the reading results on the display unit 5 and finds a misread character, he uses the keyboard 6 to move the cursor to the position of the misread character. Next, when the operator presses the candidate character selection key 61 on the keyboard 6, the code conversion circuit 7 transfers the candidate character selection command signal to the display control section 4 as described above. When the candidate character selection command signal is input, the display control unit 4 sends another candidate character category corresponding to the same character image to the display unit 5, and displays a character pattern corresponding to the candidate character pattern on the display unit 5. .

If the candidate character pattern displayed on the display screen of the display section 5 is correct, the operator presses the correction key 62 on the keyboard 6, thereby transmitting a correction command signal to the recognition control section 3. As a result, the character category is corrected.

In the above embodiment, the function of the code conversion circuit 7 has been explained as being separate from the recognition control section 3, but it can also be incorporated into the recognition control section 3.

〔Effect of the invention〕

As described above, according to the present invention, it is possible to improve the efficiency of correcting OCR recognition results, and the overall processing capacity of Uni-OCR is improved.

[Brief explanation of drawings]

FIG. 1 is a functional block diagram of an embodiment of the present invention. l...0CR 2...Reading character storage buffer 3...Recognition control unit 4...Display control unit 5...Display unit 6...Keyboard 7... Code conversion circuit

Claims (1)

[Claims] (1) In a character reading correction device that has a candidate character storage memory and corrects character reading results through operator intervention, a character category string consisting of a plurality of character categories obtained from a plurality of character images using the character reading device. and a character category obtained by using a character reading device from a character image obtained by synthesizing the plurality of character images in the candidate character storage memory; 1. A character reading and correcting device comprising means for correcting a character category by selecting either a candidate character category or a candidate character category string in a candidate character storage memory.