• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.33B no.3
• /
• pp.176-185
• /
• 1996

Transformin the huge size of drawings into a suitable format for CAD system and recognizng the contents of drawings are the major concerans in the automated analysis of engineering drawings. This paper proposes some methods for text/graphics separation, symbol extraction, vectorization and symbol recognition with the object of applying them to electronic cirucit drawings. We use MBR (Minimum bounding rectangle) and size of isolated region on the drawings for separating text and graphic regions. Characteristics parameters such as the number of pixels, the length of circular constant and the degree of round shape are used for extracting loop symbols and geometric structures for non-loop symbols. To recognize symbols, nearest netighbor between FD (foruier descriptor) of extractd symbols and these of classification reference symbols is used. Experimental results show that the proposed method can generate compact vector representation of extracted symbols and perform the scale change and rotation of extracted symbol using symbol vectorization. Also we achieve an efficient searching of circuit drawings.

• MALSORI
• /
• no.46
• /
• pp.117-126
• /
• 2003

This paper presents our language model adaptation for Korean spontaneous speech recognition. Korean spontaneous speech is observed various characteristics of content and style such as filled pauses, word omission, and contraction as compared with the written text corpus. Our approaches focus on improving the estimation of domain-dependent n-gram models by relevance weighting out-of-domain text data, where style is represented by n-gram based tf/sup */idf similarity. In addition to relevance weighting, we use disfluencies as Predictor to the neighboring words. The best result reduces 9.7% word error rate relatively and shows that n-gram based relevance weighting reflects style difference greatly and disfluencies are good predictor also.

• Korean Journal of Cognitive Science
• /
• v.1 no.1
• /
• pp.77-102
• /
• 1989

This paper reports on the development of a recognition system for Korean character,numbers and punctuation marks by syntactic approach after extracting a character or punctuation mark from a page of text.First,using the projection profile(Masudaet.al.1985,Pavlidin 1981)method, we segment a page into different regions of column or row major and then extracts lines of characters from it.Considering the height,width and connectivity of character block,we proceed to extract syllables from the extracted lines.Basically we distinguish syables into six types of formal pattern(남궁재찬 1982,이주근등 1981)following the research of lee and others,and the punctuation marks and numbers into two kinds of formal patterns,and discriminate the surface structure of the extracted syllables.By Index-Removal algorithm,we subdivide them into 44 kinds of basic korean subpattern and special characters (numbers,punctuation marks)and recognize them by syntactic method(이주근등 1981.)

• MALSORI
• /
• no.56
• /
• pp.207-223
• /
• 2005

The goal of language model adaptation is to improve the background language model with a relatively small adaptation corpus. This study presents a language model adaptation technique where additional text data for the adaptation do not exist. We propose the information retrieval (IR) technique with N-gram language modeling to collect the adaptation corpus from baseline text data. We also propose to use a dynamic language model interpolation coefficient to combine the background language model and the adapted language model. The interpolation coefficient is estimated from the word hypotheses obtained by segmenting the input speech data reserved for held-out validation data. This allows the final adapted model to improve the performance of the background model consistently The proposed approach reduces the word error rate by $13.6\%$ relative to baseline 4-gram for two-hour broadcast news speech recognition.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.57 no.10
• /
• pp.1883-1887
• /
• 2008

Most of digital media contents are composed with video and audio, picture and animation informations. Sometime, there is some deviation of information recognition quality for the video and audio information according to information receiver's characteristics or the understanding. But visual information using the text provide most clear and accurate ways for information recognition to human being. In this paper, we propose a new broadcasting system(BSDMC) to transmit clear and accurate meaning of the digital media contents. We implement general-purpose components to display the video, picture, text and symbol simultaneously. Only plug-in and call these components with proper parameters on the application developing tool, we can easily develop the multimedia contents broadcasting system. These components are implemented based on the object-oriented framework and modular structure so that increase the reusability and can be develop other applications quick and reliable.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.2
• /
• pp.722-733
• /
• 2012

Recently various studies on various applications using images taken by mobile phone cameras have been actively conducted. This study proposes a correction method for misrecognition of Korean Texts in signboard images using improved Levenshtein metric. The proposed method calculates distances of five recognized candidates and detects the best match texts from signboard text database. For verifying the efficiency of the proposed method, a database dictionary is built using 1.3 million words of nationwide signboard through removing duplicated words. We compared the proposed method to Levenshtein Metric which is one of representative text string comparison algorithms. As a result, the proposed method based on improved Levenshtein metric represents an improvement in recognition rates 31.5% on average compared to that of conventional methods.

• Speech Sciences
• /
• v.2
• /
• pp.57-66
• /
• 1997

This paper studies the effects of the method, CMS(Cepstral Mean Subtraction), (which compensates for some of the speech distortion. caused by telephone channels), on the performance of the text-dependent speaker identification system. This system is based on the VQ(Vector Quantization) and HMM(Hidden Markov Model) method and chooses the LPC-Cepstrum and Mel-Cepstrum as the feature vectors extracted from the speech data transmitted through telephone channels. Accordingly, we can compare the correct recognition rates of the speaker identification system between the use of LPC-Cepstrum and Mel-Cepstrum. Finally, from the experiment results table, it is found that the Mel-Cepstrum parameter is proven to be superior to the LPC-Cepstrum and that recognition performance improves by about 10% when compensating for telephone channel using the CMS.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2016.11a
• /
• pp.194-196
• /
• 2016

본 논문에서는 안드로이드 환경에서 글자 인식을 위한 전처리 과정으로 입력 영상에서 글자 영역만을 추출하는 기법을 제안한다. 대부분의 글자 인식 어플리케이션에서 글자를 인식하는 방법은 RoI(Region of Interest)에 인식하려는 글자를 위치시켜 놓고 사용자가 촬영함으로써 진행된다. 하지만 촬영된 영상 그대로를 인식에 사용하기 때문에 잡음 및 글자가 아닌 영역들을 글자로 인식하는 문제 등으로 인하여 인식률이 현저히 떨어진다. 제안하는 기법에서는 MSER(Maximally Stable Extremal Regions) 기법을 통해 각각의 글자를 추출한 후, 글자의 특성을 이용하여 글자 영역만을 추출한다. 기법의 성능 평가는 무료 OCR(Optical Character Recognition) 엔진인 Tesseract-OCR을 통해 글자 인식률을 비교하였으며, 제안하는 기법을 적용한 글자 인식 시스템이 적용하지 않은 시스템보다 글자의 인식률이 향상되는 것을 확인하였다.

• Proceedings of the KSPS conference
• /
• 2003.05a
• /
• pp.83-86
• /
• 2003

This paper presents our style-based language model adaptation for Korean conversational speech recognition. Korean conversational speech is observed various characteristics of content and style such as filled pauses, word omission, and contraction as compared with the written text corpora. For style-based language model adaptation, we report two approaches. Our approaches focus on improving the estimation of domain-dependent n-gram models by relevance weighting out-of-domain text data, where style is represented by n-gram based tf*idf similarity. In addition to relevance weighting, we use disfluencies as predictor to the neighboring words. The best result reduces 6.5％ word error rate absolutely and shows that n-gram based relevance weighting reflects style difference greatly and disfluencies are good predictor.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.10a
• /
• pp.89-92
• /
• 2016

The latest the siri and like this is offering services that recognize and respond to words in the smartphone or web services. In order to handle intelligently these voices, It needs to search big data in the cloud and requires the implementation of parsing context accuracy given. In this paper, I would like to propose the study on the intelligent personal assistant development method base on the Open source with ASR(Automatic Speech Recognition), QAS(Question Answering System) and TTS(Text To Speech).