JPS588382A - Character reader - Google Patents

Abstract

PURPOSE:To make the career of a file clear and to make the retrieval of mis- reading easy, by setting the reading mode to the 1st and 2nd modes and storing all the reading information at one mode and the information only corresponding to reading impossible characters at the other mode to a log file. CONSTITUTION:A recording medium 1 is scanned at a reading section 2 and the pattern of the result of scanning is stored in a reading information buffer 9. A recognizing section 3 is operated when one line of this pattern is finished, compared with a reference pattern stored in advance, the code of character pattern in coincidence to a reading character code buffer 4 and when the character is illegible, the code corresponding to the character is stored. Next, the illegible character code in the buffer 4 is displayed on a display 6 with an illegible character processing section 5. According to the 1st and 2nd reading mode set on the mode selector 10, all the reading mode is stored in a data output floppy 8 at the 1st mode and only the information of reading impossible character at the second mode is stored to a reading information buffer 12.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a character reading device, and particularly to a character reading device that is provided with a logging file to store image information of read characters.

In recent years, character reading devices have been developed and are widely used as devices for inputting characters written on letters or forms into computers. Then, by scanning the characters on the form, read information is created in @0#l"11' pattern, and the read information is compared with pre-stored character information in the character recognition unit, and the corresponding character code is The processing unit is usually one line of the form, and if there is an unreadable line in the form,
In other words, if there is a character that cannot be recognized, reading information (image information of the character) is displayed on the display connected to the character reading device.
is displayed, and the operator visually judges it and inputs it from the keyboard. - At the next stage after the character processing is completed, the stored character information, which is the image information of the characters, is discarded, and the character information of the next line is stored. Of course, recognized character codes are output as reading results.

However, in the above-mentioned character reading device, even if characters that are difficult to read (unreadable characters) or characters that are read incorrectly (misread characters) occur after reading all of the form data once, for some reason they may not be read or misread later. It's easy to find out who your new friends are. Even characters written by the same Onijin may become illegible depending on the day, and - characters with a font that is further away from Kensho semi-characters are reliably read, and - characters with a font closer to Kensho characters are readable. Misreading or misreading may occur, and it is not possible to investigate the cause. Furthermore, it has not been possible to create statistical data on the tendency of characters to be unreadable or misread.

The purpose of this invention is to overcome the shortcomings of durability, and this purpose is to set at least 2 reading modes in a character reading device that reads all characters on a recording medium and performs character recognition based on this reading information. In the first mode, all the above read information is stored in the log file, and in the second mode, only the 9 information corresponding to the characters that could not be recognized as a result of the above recognition among the above read information is stored in the log file. This is achieved by a character reading device that can be stored in a file and can arbitrarily select the Mukuro mode. Hereinafter, the present invention will be explained in detail using the drawings.

Fig. 1 is an embodiment of the present invention, Fig. M2 is a diagram showing screen display money, in the figure l is a form, 2 is a reading section, 3 is a recognition section,
4 is a readable character code buffer which is the recognition result of the familiarization unit 3, 5 is an unreadable character processing unit, 6 is a display, 7 is a data output unit, 8 is an output floppy that outputs the readable character code, and 9 is a character a read information buffer that stores all the read information that is the read output of the scanner; 10 is a mode select section;
11 outputs all the contents of the read information buffer 9 to the log floppy 12 when the mode select section 10 indicates mode 1-, and when mode 2t- indicates that there are unreadable characters, the operation will be explained next. The form 1 is scanned by a scanner provided in the reading unit 2, and the scan results for one line are, for example, "40" and "1".
” pattern is stored in the read information buffer 9. When one line ends, the g-text section 3 operates and sequentially compares with the pre-stored standard character patterns, and reads the characters corresponding to the matching pattern. The code is stored in the read character code buffer 4.

At this time, if there are any unreadable characters, a code indicating the unreadable characters is stored.

When all the processing of the U recognition section 3 regarding the contents of the read information buffer 9 is completed, the unreadable character processing section 5 starts operating. The unreadable character processing unit 5 takes out reading information corresponding to the unreadable character code in the readable character code string ν·fa 4 from the reading information buffer 9 and displays it on the display 6. In the embodiment, when an unreadable character exists, the contents of the readable character code buffer are displayed on the display 6, and the readable information corresponding to the character pointed at by the cursor 21 is displayed in a partial area 22 of the display screen. There is. (Display character pattern 23)
When a character is specified using a keyboard or the like, a code corresponding to the character is stored in the read character code buffer 4. The contents of the read code buffer 4 are then stored in the output floppy 8 by the data output section 7.

In synchronization with the operation of the data output unit 7, the log output unit 11 detects that when the mode select unit IO indicates mode 1,
The contents of the read information buffer 9 are output to the log floppy 12. If the seventh select section 10 is in mode! , the reading code buffer 4 is checked and the reading information 2 o regarding unreadable characters is output to the 70 thread 12.

Therefore, according to the present invention, by setting the mode selector 10t to mode 1, all read information, that is, all read character images are stored in the log floppy, and the reading results of the character reading device can be read by a computer. Even if a misread character is found as a result of processing, the character image corresponding to the misread character remains on the log floppy, which may cause the misreading.
-It is easy to investigate. Furthermore, by setting the mode select section 10'i-mode 2, only unreadable characters are stored in the 9gf floppy 12, so by analyzing this content, it is possible to know the tendency of unreadable characters. can also be done easily.

In the example described above, a log floppy was used as a medium for outputting read information, but it is of course possible to use other magnetic disk devices or magnetic tape devices as well, and the media installed in these devices can be updated one by one. The same medium may be used repeatedly.

Of course, instead of using a separate device as the output floppy disk, it is also possible to use the same device as the medium that outputs the read information.

[Brief explanation of the drawing]

FIG. 1 shows an embodiment of the present invention, and FIG. 2 shows a display screen of a display. In the figure, 1 is a recording medium, 2 is a reading section, 3 is a recognition section,
5 is an unreadable character processing section, 6 is a display, 8 is a data output floppy disk, 9 is a read information buffer, 10 is a mode select section, and 12 is a read information buffer.

Claims (1)

[Claims] In a character reading device that reads characters on a ML recording medium and simulates the characters based on this reading information, at least two reading modes are provided, and in the mode j[1...
All of the above reading information is stored in the log file, and in the second monido, only the reading failure information corresponding to the characters that could not be recognized is stored in the log file as a result of the above recognition organization among the reading information. Character reading device with all features that mode gold can be selected arbitrarily.