JPS61107486A - Character recognition post-processing system - Google Patents

Abstract

PURPOSE:To improve identification ratio of solid-written character string by searching candidate characters for preregistered specific characters sequentially, and segmenting an input character string at the position of that specific characters. CONSTITUTION:A character input division 2 scans an input sheet 1 by an OCR, and sends information of strength of light to an identification division 3. The identification division 3 selects candidate characters for input characters, coordinates the selected candidate characters in the order of smaller degree of difference, and stores them in a candidate memory 4. If the input character string is an address, for example, specific characters such as prefecture, city, town, and village are registered and stored in advance in a specific character registration division 5 according to the levels in the orders appearing in the input character string. When the candidate character string is stored in the candidate memory 4, a specific character retrieval division 7 refers to the specific character registration division 5, and checks if specific characters appears in the candidate characters in the order from the beginning. If the specific characters are detected, a candidate character string abstraction division 8 cuts out the character string of that part.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a character recognition post-processing method, in particular, to a pattern-recognized candidate character, focusing on whether or not there is a specific character among the candidate characters. This invention relates to a character recognition post-processing method in which post-processing is performed by finding the break in one character string and comparing the extracted character string with a dictionary.

(Prior Art and Problems) FIG. 3 shows a diagram for explaining problems with the conventional method.

Conventional 1 For example, in order to increase the recognition rate in a character recognition device that recognizes Chinese characters by optical means.

For example, post-processing is performed to check against a specific knowledge dictionary such as an address dictionary. According to the conventional method, for example, when entering an address on the 9-person sheet 1 as shown in Figure 3, the character string to be entered for each of 3 prefectures, 1 city, 1 ward, town, and village must be entered in 5 sections for post-processing. The problem was that I couldn't do it. In other words, in order to perform post-processing, it was necessary to use a format paper that had input frames divided in advance into prefecture input frames, city/ward input frames, 1 ward/town/village input frames, etc. in input sheet 1. .

As a result, the number of slots for one person to work increases, making it difficult to format the paper, making it difficult to fill in information, and making it difficult to read the written characters.

[Means for solving problems]

The present invention aims to solve the above-mentioned problems, and performs post-recognition processing on a character string written in solid form (for example, in the case of an address, a character string written without separating prefectures, cities, wards, etc.). To provide a character recognition post-processing method that increases the processing rate. Therefore, in the character recognition post-processing method of the present invention, a candidate character is output for each character as a recognition result of an input character string, and a character recognition device performs post-processing to select a final candidate for those candidate characters. In character recognition post-processing methods.

A specific character registration unit that registers and records specific characters that delimit the input character string in advance; and a dictionary that stores meaningful terms corresponding to the level determined by the specific characters for the character strings delimited by the specific characters. (5) a specific character search unit that searches the candidate characters and finds the position of the specific character in the candidate characters; and (5) cutting out a character string from the candidate characters based on the position information of the specific character obtained by the specific character search unit. The present invention is characterized by comprising a candidate character string extraction section and a dictionary matching section that compares the character string extracted by the candidate character string extraction section with the dictionary.

[Effect]

The present invention utilizes the fact that a specific input character exists among candidate characters as a result of recognizing input characters.

For example, in the case of an address, we sequentially search to see if a specific character registered in advance, such as 1 prefecture, city, 9 wards, town, etc., exists among the candidate characters, and change the input character string to 4. . . □I□, I. Yo6. It is possible to perform post-recognition processing on character strings such as yoi, -, t, nee@-', and 1.

〔Example〕

Embodiments will be described below with reference to two drawings.

FIG. 1 shows the configuration of an embodiment of the present invention, and FIG. 2 is a diagram for explaining the processing aspect of the embodiment of the present invention.

In the figure, 1 is an input sheet, 2 is a character input unit such as OCR, 3 is a recognition unit that analyzes patterns of input characters and selects candidate characters, 4 is a candidate memory in which the selected candidate character strings are stored, and 5 6 is an address dictionary; 7 is a specific character search unit that searches for specific characters in the candidate memory 4;
9 is a candidate character string extraction unit that extracts candidate character strings delimited by specific characters; 9 is a dictionary collation unit that collates the candidate character strings extracted by the candidate character string extraction unit 8 with the address dictionary 6; 10 is a dictionary collation result. Represents the result output section to be output. Also, in FIG.

Reference numeral 20 represents an input character string.

Case 2 of the present invention For example, as shown in the character string 20 shown in FIG. 2, the address can be written in solid text on the manual sheet 1 without dividing it into prefectures, cities, wards, etc. The character input section 2 scans the input sheet 1 by optical means such as OCR, and transmits light intensity information to the recognition section 3. The recognition unit 3 selects candidate characters for the input characters by, for example, extracting topological features or performing stroke analysis on the input information, and nine different methods are known for this recognition process. , and detailed explanation will be omitted.

For each selected candidate character, the recognition unit 3
As shown in the figure, the candidates are ranked in descending order of so-called difference degree and stored in the candidate memory 4 from the 1st to the 20th place. Needless to say, the first candidate character is not necessarily the correct input character.

By the way, when the input character string is an address, the address usually includes specific characters such as "city, prefecture, prefecture, city, county, two wards, town, village." In the specific character registration section 5, these specific characters are registered and stored in advance in correspondence with the level according to the order in which they appear in the input character string. Address 2 For example, “Miyako.

Each kanji for "Michi, 1 prefecture" is registered at the prefecture level, each kanji for "city, 1 ward" is registered at the city/ward level, and each kanji for "ku, town, village" is registered at the ward/town/village level. Registered as . Further, the specific character registration section 5 has information corresponding to each level or each specific character.

Index information is set for a dictionary that stores character strings that can appear at that level.

When the candidate character string is stored in the candidate memory 4, the specific character search unit 7 refers to the specific character registration unit 5.

Using a specific character corresponding to each level as a key, the system sequentially tests from the beginning whether or not the specific character that serves as the key appears among the candidate characters. In the example shown in FIG. 2, the character ``Miyako'' is found in the first position of the third character. As a result, it can be seen that the first to third characters are character strings at the prefecture level. Next, by sequentially starting from the 4th character, "city" will appear at the 6th character, so the 3 characters from the 4th character to the 6th character are at the city/ward level. It is recognized as

Similarly, at 7, it is recognized that the 7th to 10th characters are the word ``ku, town, village, and bell''.

Note that these specific characters do not necessarily have to appear first in the candidate ranking.1 For example, for locations where specific characters are likely to appear, such as the 3rd character or the 10th character, the candidate rankings are determined from the highest candidate ranking. Search results are selected in order.

When the specific character search unit 7 detects the position of a specific character, the candidate character string extraction unit 8 extracts a partial candidate character string from the candidate character string stored in the candidate memo IJ 4 until the specific character appears. Cut out.

The dictionary checking unit 9 is notified.

The dictionary matching unit 9 determines that the notified character string is, for example, ``prefecture''.
to)'7::b6. u@・Address Zhanshinsho 60 Dictionary 4 copies
; , , The combination of each word registered in i and the characters according to the candidate ranking in the notified partial candidate character strings is sequentially tested to see if it will fail or not. As a result, in the example shown in Figure 2, at the prefecture level, the word "
It turns out that it is Tokyo. In addition.

If there is a word that loses 2 or more in the comparison with the address dictionary 46, by calculating the points of the candidate ranking.

The character combination with the higher candidate ranking is selected.

Similarly, in the example shown in FIG. 2, at the city/town/ward level, the partial candidate character string from the 4th character to the 6th character is compared with the dictionary B part of the address dictionary 6, and "Machidafu" is determined. Obtained as a matching result.

Furthermore, at the ward/town/village level, "Shinkojimachi" is obtained as a matching result based on the partial candidate character string from the 7th character to the 10th character.

The matching result is notified to the result output unit 10, and the result output unit 1
0 confirms with the input person as necessary, determines the final recognition result, and executes output processing to a predetermined device.

The above description has been made using character input of an address as an example, but the present invention can also be applied to input of, for example, affiliation in a company.

This method can be similarly applied to cases where there are characters that often appear at the end of one character string, such as "department" or "section" as specific characters. Also.

It can be applied not only to handwritten characters but also to recognition of printed characters.

〔Effect of the invention〕

As explained above, according to the present invention, a solid input character string can be divided into word units that can be post-processed, so post-processing is performed to select the most appropriate final candidate from a plurality of candidates. It is possible.

The recognition rate can be improved. Since input character strings can be written in solid text, they are easy to write (and the entered character strings are easy to read).

Additionally, paper waste can be reduced. The inputter does not need to be aware of specific characters, and there is no burden on the inputter.

[Brief explanation of drawings]

Figure 1 shows the configuration of an embodiment of the present invention, Figure 2 is a diagram for explaining the processing mode of an embodiment of the present invention, and Figure 3 is a diagram for explaining problems with the conventional method. . In the figure, 1 is an input sheet, 2 is a character input section, 3 is a recognition section,
4 is a candidate memory, and 5 is a specific character registration section. 6 is an address dictionary, 7 is a specific character search section, 8 is a candidate character string extraction section, 9 is a dictionary collation section, and 10 is a result output section. 20 represents an input character string. Patent applicant: Fujitsu Ltd. Representative Patent Attorney Hiroshi Mori 1) (1 other person) No. 1121

Claims (1)

[Claims] In a character recognition post-processing method in a character recognition device that outputs candidate characters for each character as a recognition result of an input character string, and performs post-processing to select a final candidate for those candidate characters, the above input character string is a specific character registration unit that registers and stores specific characters that serve as delimiters; a dictionary that stores meaningful terms corresponding to the level determined by the specific characters for character strings delimited by the specific characters; and a dictionary that searches for the candidate characters. a specific character search unit that finds the position of the specific character in the candidate characters; a candidate character string extraction unit that extracts a character string from the candidate characters based on the position information of the specific character obtained by the specific character search unit; A character recognition post-processing method comprising: a dictionary matching unit that matches the character string extracted by the candidate character string extraction unit with the dictionary.