JPS6162986A - Recognition order determining system - Google Patents

Abstract

PURPOSE:To improve the recognition rate of the titled system by determining priority levels corresponding to a learning stage for individual character type categories and determining the recognition order of candidate character type categories on the basis of calculated degrees of resemblance and these priority levels. CONSTITUTION:A binary character pattern from a character observing part 1 is inputted to a pattern matching part 4 through a feature extracting part 2,a nd degrees of resemblance between a feature pattern and standard feature patterns are calculated. Character type categories in a dictionary part 3 re divided in accordance with learning stages of learning education, and priority levels are determined for individual category sets. Operating parts 4a-4d in the pattern matching part 4 add priority levels of category sets to character type codes and degrees of resemblance and store them in a candidate memory 5. A knowledge part 6 determines the recognition order of candidate character type categories on the basis of calculated degrees of resemblance and priority levels and stores character type codes in a result memory 8 in the recognition order.

Description

DETAILED DESCRIPTION OF THE INVENTION (a) Field of Industrial Application The present invention relates to a character recognition system for recognizing handwritten Chinese characters, and more particularly to a recognition ranking determination method for candidate character type categories.

(b) Conventional technology Generally, in a character recognition system, the degree of similarity between a feature pattern extracted from an input character pattern and a standard feature pattern for each character type category registered in advance in a dictionary section is calculated,
Select n1 candidate character type categories with a high degree of similarity. Then, the candidate character type category with the highest degree of similarity is output as the recognition result, and in order to correct misrecognition,
Among the selected candidate character categories, 1! In step ii, the recognition ranking from first place to first place is determined.

However, as mentioned above, using only Chijij and Ichijo to determine the recognition R ranking causes many false positives, so some kind of post-processing is performed after selecting multiple candidate character type categories based on similarity. A method for determining recognition rankings has become possible.

Conventionally, as post-processing, Japanese Patent Application Laid-Open No. 59-3208
As disclosed in Publication No. 2, there are methods that perform grammatical processing, and methods that determine whether the characters before and after the recognized character are kanji, katakana, or hiragana, as in Japanese Patent Application Laid-open No. 59-27381. Something was proposed.

(c) Problems to be solved by the invention In the conventional technology, grammatical processing is performed as post-processing, so the knowledge section such as a grammatical dictionary becomes enormous, and furthermore, the processing content becomes extremely complex. There is a problem that
Furthermore, in the method of determining whether the preceding and succeeding characters are kanji or katakana, no improvement in recognition rate could be expected if the selected candidate character categories were only kanji or hiragana.

Furthermore, difficult character categories have many characteristics, so
Easy to distinguish from other glyph categories, and there are relatively few misrecognitions, but because simple glyph categories have few characteristics, many candidate glyph categories with close similarity are selected.1. However, there were many cases where difficult character categories were ranked higher in the recognition ranking and were misrecognized.

B) Means for Solving the Problems The present invention sets a priority level for each character type category according to the learning stage of school education, and then sets a priority level for each character type category based on the degree of similarity obtained by calculation and this priority level. This determines the recognition ranking of candidate character type categories.

(e) Effects According to the present invention, the recognition ranking of the candidate character type categories that have the highest degree of similarity and have a high priority according to the learning stage of school education will be placed at the top, making it easy for the general public to recognize YO (known character type category IJ- will now have priority.

Furthermore, it is sufficient to store the relationship between similarity and priority in the knowledge section.

(F) Embodiment FIG. 1 is a block diagram of a character recognition system to which the present invention is applied, and (1) reads characters written on an input manuscript and outputs the reading result as a binary character pattern. Character observation section, (2) is a feature extraction section that extracts feature patterns from input character patterns, (3) is a character observation section that stores standard feature patterns for each character category, and (4) is a feature extraction section that extracts feature patterns from input character patterns. This is a pattern matching unit that performs matching with a standard feature pattern and calculates the degree of similarity between both patterns.

The character categories in the dictionary section (3) are divided into categories according to the learning stage of school education, and a priority is set for each category cent. That is, as shown in Figure 2, the characters learned in the first to third years of elementary school are categorized as Category 7) 1 (3a), the characters learned in the fourth to sixth years of elementary school as Category Set 2 (3b), and the characters learned in junior high school and above are categorized as Category Set 1 (3a). All character categories are divided into three category sets, such as character to be learned in category set 3 (3c), and priority levels 1 to 3 are assigned to category sets 1 to 3 in order.

The pattern matching section (4) selects category sets 1~
3, each of which selects nine candidate character type categories from each category set in descending order of similarity. The code and the similarity as a calculation result are stored in a candidate memory (5). At this time, the calculation unit adds the priority of the corresponding category set to the character type code and the similarity f, and these three pieces of information are stored in the candidate memo as information of each candidate character type category. In this way, a total of 3rL candidate character type categories, nine of each category cent, are stored in the candidate memory (5).

The knowledge section (6) stores the relationship between similarity and priority, and the clustering control processing section (7) refers to the contents of this knowledge section (6) and stores it in the candidate memory (5). The recognition rankings of the top nine of the 3rL candidate character type categories are determined, and the character type codes are stored in the result memo (January 8) in order of recognition ranking. For example, if we use the city block distance DI as the degree of similarity, and assume that the smaller this distance is, the greater the degree of similarity, then the knowledge section (6) specifically contains the distance DI and the priority, as shown in Figure 3. The relationship as to whether or not the recognition order can be swapped is stored.

The manner in which the recognition ranking is determined will be explained below using a specific example.

First, the candidate character type category with character type code Ml, city block distance and DI priority P (1=a, b, c-1P=1, and 3) is expressed as (M2S DI, p). For example, as shown in FIG. 4(a), the top five candidate character type categories with the highest similarity are selected from each of category sets 1 to 3, and the relationship between the city block distances Di is expressed as Dv( :D.

Assume that <D-<DQ <Db <Dr <D W.

If, as in the past, only the size of the city block distance is used, the recognition order will be as shown in FIG. 4 (b).

Now, the threshold in the knowledge section (6) is D...
The relationship between D and the calculated city block distance is D.

<Dv, Da <Dt <DQ, Dw <Da, the clustering control processing unit (7)
In order to rank the character type codes Mv, MP, M, as rank B, and the character type codes M, , Mb, Mr, Mw as rank C, and change the recognition order based on priority within these ranks. , As shown in FIG. 4(c), within each rank, character type codes with higher priority are placed at the top, and as a result, recognition rankings are determined based on similarity and priority. Then, the top 5 character type codes M@, Mp
, My , Mb.

M, are stored in the result memo (January 8) in order, and the answer output control unit (9) sends the first character type code M, to a character display device such as a word processor or a personal computer as a recognition result, and displays that character type. is displayed. At this time, if there is a misrecognition, the character type code of the next recognition rank is sent out, and thereafter, the character type codes of the next rank are sent out in sequence until a correct recognition result is obtained.

Therefore, priority is given to candidate character type categories that have a high degree of similarity and are at a low learning stage in school education, and characters that are simple and well-known to the general public are ranked high in the recognition ranking, improving the recognition rate.

Also, as another example, the city block distance Dv is sufficiently small compared to other candidates such that Dv<D, and D,<D, ,
Dt)<Dt, the character type code Mv is ranked with the A rank and the recognition order is not changed, and the character type codes M, , M, , MQ, Mb of the B rank are changed in the ranking based on the priority. (The recognition order is shown in Figure 4)
become that way. In this case as well, when compared to FIG. 4 ((b)), the recognition ranking of the candidate character type category with a low learning stage comes to the top.

Therefore, in this embodiment, the character type categories are divided in advance into category sets according to the priorities, and the priority information is obtained. It may be stored in advance.

Assigning priorities according to the learning stage can be easily done by referring to the Ministry of Education's elementary school guidance manual and junior high school guidance manual.

(g) Effects of the Invention According to the present invention, priority is given to character type categories that are simple and well-known to the general public, which improves the recognition rate and eliminates the need for a huge knowledge department. , the recognition ranking can be determined in a short processing time.

In addition, in particular, since priority is used according to the learning stage of school education, it is possible to provide a film-like character, which is extremely effective when using a character recognition device as an input means to a word processor or personal computer. .

[Brief explanation of the drawing]

Figure 1 is a block diagram of a character recognition system to which the present invention is applied, Figure 2 is an explanatory diagram showing the contents of the category set, Figure 3 is an explanatory diagram showing the contents of the knowledge section, and Figures 4 (a) to )
FIG. 2 is an explanatory diagram showing a specific example of recognition ranking determination. Explanation of main figure numbers (1)... Literary view side part, (2)... Feature extraction part,
(3)...Dictionary section, (4)...Pattern munching section, (5)...Candidate memory, (6)...Knowledge section, (7)...Clustering control processing section, ( 8)
...Result memory, (9)...Answer output control section.

Claims (1)

[Claims] (1) A feature pattern extracted from the input character pattern,
In a character recognition system that selects a plurality of candidate character categories by calculating the degree of similarity with standard feature patterns for each character type category registered in advance in the dictionary section, each character type category is assigned a learning stage in school education. A recognition order determining method, characterized in that a priority is determined according to the above, and the recognition order of the plurality of candidate character type categories is determined based on the similarity obtained by the calculation and the priority.