• 한국콘텐츠학회논문지
• /
• 제18권7호
• /
• pp.495-502
• /
• 2018

딥 러닝 기술의 등장으로 여러 나라의 필기체 인식은 높은 정확도 (중국어 필기체 인식은 97.2%, 일본어 필기체 인식은 99.53%)를 보인다. 하지만 한글 필기체는 한글의 특성으로 유사글자가 많은데 비해 문자의 데이터 수는 적어 글자 인식에 어려움이 있다. 하이브리드 러닝을 통한 한글 필기체 인식에서는 lenet을 기반으로 하여 낮은 레이어를 가진 모델을 사용하여 한글 필기체 데이터베이스 PE92에서 96.34%의 정확도를 보여주었다. 본 논문에서는 하이브리드 러닝에서 사용하였던 데이터 확장 기법(data augmentation)이나 multitasking을 사용하지 않고도 GoogLenet 네트워크를 기본으로 한글 필기체 데이터에 적합한 더 깊고 더 넓은 CNN(Convolution Neural Network) 네트워크를 도입하여 PE92 데이터베이스에서 98.64%의 정확도를 얻었다.

• 한국인쇄학회지
• /
• 제6권1호
• /
• pp.41-57
• /
• 1988

Recently, it Is required to develop OCR(Optical Character Reader) along with the progress of the information processing system for Hangeul. Characters have to be recognized clearly so that OCR can be applied, Structure analysis method and lump method are used for the recognition of characters, and OCR is now available for the recognition of printed characters and handwritten alphanumeric characters having simple structure by them However, It is known that there should be much more study on the development of handwritten Hangout's OCR. This paper proposed a new method for the handwritten Hangout character recognition. The units of Initial consonant of Hangout are separated and then recognized from the utilization of the position- Information of Hangeul's units from the normalized patterns using the regression line theory. It is carried out for the extraction of the block which exists in the virtual Initial consonant region from the normalized input patterns and the calculation on maximum value (${\beta}$) of likelihood after comparing the features of separated subpattern with the initial consonant dictionary.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• 제27권2호
• /
• pp.135-145
• /
• 2023

Recently, as deep learning technology has developed, various deep learning technologies have been introduced in handwritten recognition, greatly contributing to performance improvement. The recognition accuracy of handwritten Hangeul recognition has also improved significantly, but prior research has focused on recognizing 520 Hangul characters or 2,350 Hangul characters using SERI95 data or PE92 data. In the past, most of the expressions were possible with 2,350 Hangul characters, but as globalization progresses and information and communication technology develops, there are many cases where various foreign words need to be expressed in Hangul. In this paper, we propose a model that recognizes and combines the consonants, medial vowels, and final consonants of a Korean syllable using a multi-label classification model, and achieves a high recognition accuracy of 98.38% as a result of learning with the public data of Korean handwritten characters, PE92. In addition, this model learned only 2,350 Hangul characters, but can recognize the characters which is not included in the 2,350 Hangul characters

• 대한전자공학회논문지
• /
• 제24권5호
• /
• pp.769-776
• /
• 1987

A method of on-line recognition of Korean characters (Hangeul) by augmented conterxt free grammar is described in this paper. Syntactic analysis with context free grammar oftern has ambiguity. Insufficient description of relations among Hangrul sub-patterns causes this ambiguity can be determined through repetition of experiments. Flexible syntactic analysis is executed by adapting the condition to the (advice)part of augmented context free grammar. The ratio of correct recognition of this method is more than 99%.

• 전자공학회논문지B
• /
• 제28B권10호
• /
• pp.747-754
• /
• 1991

In this paper, we studies the on-line recognition of handwritten character using WLD(weighted levenshtein distance) algorithm with parallelism. The Hangeul can be separated for unit of phonemes and the alphanumeric can be separated for unit of characters. And, we studies the parallelism and the concurrency of the WLD algorithm for realization of special-purpose processor. By the simulation result for 10, 000 characters in practical sentences, the recognition rate of strokes in obtained 96.57$\%$ and the separation rate for phonemes and characteristics is obtained 95.4$\%$.

• 대한전자공학회논문지
• /
• 제27권11호
• /
• pp.103-110
• /
• 1990

本 論文에서는 퍼지推論을 利用한 온라인 筆記體 한글文字의 認識에 관하여 硏究하였다. 筆記者 마다의 筆體變動으로 인한 애매성을 解決하기 위해, 각 스트로크間의 相對位置情報마다 作成되는 生成 을 퍼지推論에 適用하여 筆記體 한글을 認識하였다. 處理시간을 短縮하기 위하여, 入力文字의 스트로크 數에 따라 미리 分類한 基準文字의 小群을 선택하였고, 이 小群의 文字들과 入力文字와의 거리에 許容限果를 주어 基準文字들을 감소시켜 퍼지推論에 적용하였다. 10人으로부터 수집된 39990字의 筆記體 한글文字에 대하여 實驗한 結果, $99.5{\%}$의 認識과 0.4초/文字의 平均處理 速度를 얻었다.

• 대한전자공학회논문지
• /
• 제23권1호
• /
• pp.114-119
• /
• 1986

In this paper, an algorithm for the on-line recognition of handwritten Hangeul is proposed. The strokes are recognized by the minimum distance parser. The phonemes are separated by the finite-state automata resulted from the state graph of phonemes which are produced by the order of strokes. By simulation result for 3,000 characteristics in practical sentences, the recognition rate of strokes is obtained to be 98.5% and the separation rate of phonemes is obtained to be 92.5%.

• 대한전자공학회논문지
• /
• 제22권5호
• /
• pp.1-9
• /
• 1985

In this paper, a new methods which decomposes the handwritten-Hangout shape panerns into subpatterns and recognizes the decomposed subpatterns are proposed. the feature vcfices arc detected by searching boundary of the shape pattern and a topolo-gical structure is represented by a bridge links and contact links between the feature vertices. From the tpcological structure, Hangout shape patterns are decomposed into the subpatterns of 44-Korean alphabet. The 학obol and the local attributes are extracted from the subpattrrns and the subpatterns are recognized by matching those attributes with the dictionary. It is assured that this method is more effect and reasonable for deformed handwrioen Hangout shape patterns. Experimental results show that recognition rate is 99(%) and recogni-tion time is also reduced as those using the thinning process.