JPH0252315B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention particularly relates to an optical character reading device with an improved dictionary selection method.

Currently, optical character reading devices (hereinafter referred to as OCR) for input into English word processors are becoming popular. However, conventionally, the above OCR
In most cases, the types of characters that can be read are limited to OCR fonts, and there are only a few that can be used with commonly used office fonts. In addition, multi-font OCR that can read various fonts is very expensive compared to word processors, so it is difficult to maintain a good balance.
It was unsuitable for OCR. Furthermore, when using OCR for multiple fonts, it was necessary to select the font using a program or a switch.

This invention was made in view of the above circumstances, and provides an optical character reading device that can read multiple types of fonts by switching the reading dictionary based on predetermined characters on a form. The purpose is to provide.

Hereinafter, one embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a schematic diagram of an embodiment of the present invention. In the figure, reference numeral 1 indicates a photoelectric conversion section. The photoelectric conversion unit 1 has a function of photoelectrically converting characters on a form and outputting the photoelectrically converted character pattern. Reference numeral 2 indicates a recognition section. This recognition section 2 has a function of recognizing the photoelectrically converted character pattern. code 3
indicates a dictionary index table. This dictionary index table 3 stores the starting addresses of corresponding storage areas of dictionaries corresponding to various fonts. Reference numeral 4 indicates a multi-font dictionary. This multi-font dictionary 4 is composed of independent dictionaries for each of a plurality of fonts. Reference numeral 5 indicates an accumulator. The accumulator 5 is configured to store the similarity value between the standard pattern in the dictionary for each font in the multi-font dictionary 4 and the matching result for the character in the first line of the document.

Next, the operation of the above embodiment will be explained. First, as shown in Figure 2, the form used in this OCR has a specific character arrangement predetermined in the font used in the form, in this case the character string ABC, in the first line of the form. Print it out. The font types include Courier12, Courier72,
There are Prestige Elite72, Prestige Pica, etc. From the second line onwards, data and text can be freely printed in the same font.

The characters in such a form are written in the photoelectric conversion unit 1.
is photoelectrically converted. At this time, first, specific characters printed on the first line of the form are photoelectrically converted. The recognition unit 2 refers to the contents of the dictionary index table 3, reads out the standard pattern corresponding to the character printed in the first line from one dictionary in the Malmfont dictionary 4, and Matching is performed between the standard pattern and the photoelectrically converted character pattern in the first line. Then, the sum of the similarity values of the three characters in the first row, which is the result of this matching, is calculated by the accumulator 5.
is memorized. Further, the same operation as described above is repeated for each dictionary in the multi-font dictionary 4, and the sum of the similarity values of the three characters, which is the matching result for each dictionary, is sequentially stored in the accumulator 5. In this way, the accumulator 5
The recognition unit 2 compares the sum of similarity values for each dictionary in the multi-font dictionary 4 stored in the multi-font dictionary 4 . At this time, the dictionary with the largest sum of similarity values is selected from the multi-font dictionary 4.

From the second line onward, recognition is performed using the selected dictionary. In this case, if unrecognized characters occur frequently in a line in the middle, dictionary selection is performed again for that line. This selects a dictionary based on the number of unrecognizable characters and the similarity value of each character in the line.

Therefore, such OCR has the following effects.

(1) Since the characters required to select a dictionary are determined and the standard character patterns read from each dictionary are limited, the time required for dictionary selection can be reduced.

(2) Compared to creating a mix multi-dictionary for various fonts, it is easier to create a dictionary because it only requires a collection of individual dictionaries for each font, and there is no need to consider the occurrence of similar characters.

(3) Compared to MIX multi-dictionaries, it is possible to read with high accuracy and at the same time, it is possible to speed up the recognition time.

(4) The operator only needs to print the corresponding character string predetermined in the type head being used at that time on the first line of the ledger, so there is no burden on the operator.

(5) Since the multi-font dictionary 4 can be configured by combining only necessary font dictionaries, it can be easily made optional.

As described above, according to the present invention, it is possible to provide an optical character reading device that can read multiple types of fonts by switching the reading dictionary based on predetermined characters on a form. .

[Brief explanation of the drawing]

FIG. 1 is a schematic diagram of an embodiment of the present invention, and FIG. 2 is a diagram showing a form used in the embodiment. 1... Photoelectric conversion unit, 2... Recognition unit, 3... Dictionary index table, 4... Multifont dictionary, 5... Accumulator.

Claims (1)

[Claims] 1. A photoelectric conversion unit that photoelectrically converts characters on a form;
A multi-font dictionary consisting of a plurality of dictionaries corresponding to each font, and a similarity value obtained by referring to each of the plurality of dictionaries for the photoelectrically converted character pattern of a specific character at a specific position on the form. The accumulator stored for each dictionary and the similarity value are determined, and the dictionary corresponding to the highest similarity value stored in the accumulator is selected from the multi-font dictionary, and the selected dictionary is referred to for photoelectric conversion. 1. An optical character reading device comprising: recognition means for recognizing a converted character pattern.