• Proceedings of the Korean Institute of Communication Sciences Conference
• /
• 1985.10a
• /
• pp.109-112
• /
• 1985

This paper proposes a algorithm of phonemes extraction in korean character pattern recognition. The phonemes are classified into the patterns which are separable and connected with each other. The former is extracted by means of pattern matching in consideration of topological structure of ponemes and direction of stroke sequentially. The latter is extracted by means of index and window algorithm which are performed by a 3$\times$3 sequential local operation in the thinned character pattern.

• Textile Coloration and Finishing
• /
• v.29 no.4
• /
• pp.256-267
• /
• 2017

The purpose of this study was to analyze the functional ingredients of Perilla Frutescens L. Britt extracts and to confirm the possibility of producing PVA nanofibers using extracts. Distilled water, 3% aqueous sodium hydroxide solution and ethanol were used as extraction solvents. The electrospinning was carried out at a PVA concentration of 12%, an applied voltage of 10 kV and a tip to collector distance of 15cm. The contents of volatile substances, essential oils, total polyphenols and flavonoids of the extracts were measured to examine the constituents of functional materials. Flavor components and esters were identified in 3% sodium hydroxide and ethanol extracts. The content of polyphenols and flavonoids in ethanol extracts was higher than that of medicinal plants. 1wt.% of Tween 20 was added to disperse the essential oil components of the ethanol extract. Addition of a dispersant made it possible to produce a homogeneous mixture by having some compatibility with the ethanol extracts and the PVA molecule. When the concentration of the ethanol extract was 0.25 and 0.5wt%, relatively uniform PVA nanofiber having an average diameter of 350 to 365nm could be produced. The results of FT-IR, XRD and DSC analysis confirmed that Perilla Frutescens L. Britt ethanol extract was well mixed with PVA molecules and was electrospun.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.5
• /
• pp.1281-1293
• /
• 1996

In this paper, we propose a novel recognition model of on-line cursive Korean characters using HMM(Hidden Markov Model) and level building algorithm. The model is constructed as a form of recognition network with HMM for graphemes and Korean combination rules. Though the network is so flexible as to accomodate variability of input patterns, it has a problem of recognition speed caused by 11, 172 search paths. To settle the problem, we modify the level building algorithm to be adapted directly to the Korean combination rules and apply it to the model. The modified algorithm is efficient network search procedure time complexity of which depends on the number of HMMs for each grapheme, not the number of paths in the extensive recognition network. A test with 15, 000 hand written characters shows recognition rat 90% and speed of 0.72 second per character.

• Journal of Korea Multimedia Society
• /
• v.8 no.2
• /
• pp.287-295
• /
• 2005

For a quick and easy way of writing Hangul in hand-held devices, we analyzed the combination rules of Hangul alphabets that we constructed a small set of Hangul alphabets that combines a whole set of Hangul alphabets. Hangul consonants are generated from base set of consonants with function keys and vowels are constructed from eight base set of vowels by combination rules and stroke-adding rules. Especially, we adopted an easy error-correction method to resolve the inconvenience of using a delete key for vowel harmony errors. Futhermore, we diversified the input method of diphthongs for user-friendliness by minimizing the efforts of teaming vowel combination rules. We compared our method with previous methods of 'Chon-Ji-In' and 'Na-Rat-Keul' that the proposed method is better than the previous methods in input speed and error correction functionality.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.40 no.3
• /
• pp.328-333
• /
• 1995

Leaf quality and fatty acid compositions of collected perilla related genus and species germplasms such as Perilla frutescens var. japonica Hara, Perilla frutescens var. acuta Kudo, Perilla frutescens var. crispa Decaisne, Perilla frutescens var. for viridis Makino, Mosla punctulata Nakai, Mosla japonica Maxim, Mosla dianthera Maxim were analysized. The number of leaves per tiller and leaf size of perilla germplasms were more and bigger than those of mosla germplasms. Aroma degree of mosla germplasms was higher than aroma degree of perilla germplasms. Mosla germplasms could be utilized in the breeding for high aroma perilla lines. Otherwise, the softness of perilla germplasms was higher than that of mosla germplasms. In case of oil and protein contents, perilla germplasms was higher than mosla germplasms, however compositions ratio of fatty acid, especially linolenic acid of mosla germplasms was higher than that of perilla germplasms, therefore mosla germplasms could be utilized as breeding materials with high linolenic acid for industrial oil. The linolenic acid with excellent quality and unsaturated fatty acid showed negative correlation with oil content, protein content and saturated fatty acids.

• Journal of Korea Multimedia Society
• /
• v.6 no.2
• /
• pp.216-223
• /
• 2003

In this paper, we present SCCRS(Speech and Character Combined Recognition System) for speaker /writer independent. on-line multimodal interfaces. In general, it has been known that the CHMM(Continuous Hidden Markov Mode] ) is very useful method for speech recognition and on-line character recognition, respectively. In the proposed method, the same CHMM is applied to both speech and character recognition, so as to construct a combined system. For such a purpose, 115 CHMM having 3 states and 9 transitions are constructed using MLE(Maximum Likelihood Estimation) algorithm. Different features are extracted for speech and character recognition: MFCC(Mel Frequency Cepstrum Coefficient) Is used for speech in the preprocessing, while position parameter is utilized for cursive character At recognition step, the proposed SCCRS employs OPDP (One Pass Dynamic Programming), so as to be a practical combined recognition system. Experimental results show that the recognition rates for voice phoneme, voice word, cursive character grapheme, and cursive character word are 51.65%, 88.6%, 85.3%, and 85.6%, respectively, when not using any language models. It demonstrates the efficiency of the proposed system.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.33B no.11
• /
• pp.140-149
• /
• 1996

Cursive characters are written on an economical principle to reduce the motion of a pen in the limit of distinction between characters. That is, the pen is not lifted up to move for writing a next stroke, the pen is not moved at all, or connected two strokes chance their shapes to a similar and simple shape which is easy to be written. For these reasons, strokes and korean alphabets are not only easy to be changed, but also difficult to be splitted. In this paper, we propose a curvature stroke modeling method for splitting and matching by using a structural primitive. A curvature stroke is defined as a substroke which does not change its curvanture. Input strokes handwritten in a cursive style are splitted into a sequence of curvature strokes by segmenting the points which change the direction of rotation, which occur a sudden change of direction, and which occur an excessive rotation Each reference of korean alphabets is handwritten in a printed style and is saved as a sequence of curvature strikes which is generated by splitting process. And merging process is used to generate various sequences of curvature strikes for matching. Here, it is also considered that imaginary strokes can be written or omitted. By using a curvature stroke as a unit of recognition, redundant splitting points in input characters are effectively reduced and exact matching is possible by generating a reference curvature stroke, which consists of the parts of adjacent two korean alphasbets, even when the connecting points between korean alphabets are not splitted. The results showed 83.6% as recognition rate of the first candidate and 0.99sec./character (CPU clock:66MHz) as processing time.

• The Transactions of the Korea Information Processing Society
• /
• v.2 no.6
• /
• pp.839-846
• /
• 1995

In this paper, the spelling errors which are made by human being in the real word processors are collected and analyzed. Based on these data, we make a prototype which can perform spell aid function providing candidate words. The number of candidate characters are minimized by the frequency of Jaso and character, so the number of candidate words could be minimized. The average number of candidate words presented are 3.2 to 8, and 62.1 % to 84.1% of the correct words are presented in the candidate words.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1994.06c
• /
• pp.75-79
• /
• 1994

한국어 어절의 모든 가능한 형태소 분석 결과는 형태소 격자 구조로 대응된다. 즉, 형태소 분석과정은 형태소 격자 구조를 만드는 과정과 동일하다고 말할 수 있다. 기존의 방법들은 여러개의 가능한 분석 결과에 중복되는 형태소들을 그대로 저장하여 자료 관리의 비효율성이 있었다. 본 논문에서 설명하는 형태소 분석기는 형태소 분석의 중간 결과를 공유하여, 자료의 중복 저장을 피했고, 모든 가능한 형태소 분석 결과를 형태소 격자 구조의 가능한 모든 경로로 대응하였다. 한편, 형태소 배열 규칙은 품사 태깅된 말뭉치로부터 자동으로 추출되었다. 또한, 사전도 품사 태깅된 말뭉치로부터 자동으로 구축되었으며, 굴절된 형태소는 등록되지 않는다. 그러나 불규칙 및 축약 현상에 관한 정보는 수동으로 추가되었다. 불규칙 및 축약 현상의 발생 가능 위치는 한글 자소 패턴에 의해서 찾아지고, 이들 현상의 처리는 절차적인 방법에 의해 해결되었다.

• Annual Conference on Human and Language Technology
• /
• 2017.10a
• /
• pp.67-70
• /
• 2017

상호(商號)란 상인이나 회사가 영업 활동을 위해 자기를 표시하는데 쓰는 명칭을 말한다. 일반적으로 국내 기업의 상호 표기법은 한글과 로마자를 혼용함으로 상호 검색 시스템에서 단어 불일치 문제를 발생시킨다. 본 연구에서는 이러한 단어 불일치 문제를 해결하기 위해 Sequence-to-sequence 모델을 이용하여 로마자 상호를 이에 대응하는 한글 상호로 변환하고 그 후보들을 생성하는 시스템을 제안한다. 실험 결과 본 연구에서 구축한 시스템은 57.82%의 단어 정확도, 90.73%의 자소 정확도를 보였다.