• 한국컴퓨터정보학회논문지
• /
• 제15권6호
• /
• pp.73-80
• /
• 2010

어휘 인식 시스템은 부정확한 어휘 제공과 유사한 음소 인식으로 인식률이 저하되며 이는 유사한 음소인식 오인식과 효율적 특징 추출 처리를 위한 방법을 필요로 한다. 따라서 본 논문에서는 음소가 갖는 특징을 기반으로 바타챠랴 거리 측정법을 이용한 음소 유사율 오류 보정 개선 시스템을 제안하였다. 음소 유사율은 모노폰으로 훈련시킨 훈련 데이터의 음소에 HMM 특징 추출 방법을 이용하였으며 유사한 음소는 바타챠랴 거리 측정법을 이용하여 정확한 음소로 인식할 수 있도록 유도하여 인식률 향상 효과를 얻을 수 있었다. 이를 유클리디안 거리 측정법과 동적타임 워핑 시스템에 비교한 시스템 성능 평가 결과 1.2%의 향상된 97.91% 인식률을 보였다.

• 음성과학
• /
• 제13권3호
• /
• pp.155-167
• /
• 2006

Hearing impairment children possess poor underlying perceptual knowledge of the sound system and show delayed development of segmental organization of that system. The purpose of this study was to investigate the relationship between phonological awareness ability and word identification ability in hearing impaired children. 14 children with moderately severe hearing loss participated in this study. All tasks were individually administered. Phonological awareness tests consisted of syllable blending, syllable segmentation, syllable deletion, body-coda discrimination, phoneme blending, phoneme segmentation and phoneme deletion. Close-set Monosyllabic Words(12 items) and lists 1 and 2 of open-set Monosyllabic Words in EARS-K were examined for word identification. Results of this study were as follows: First, from the phonological awareness task, the close-set word identification showed a high positive correlation with the coda discrimination, phoneme blending and phoneme deletion. The open-set word identification showed a high positive correlation with phoneme blending, phoneme deletion and phoneme segmentation. Second, from the level of phonological awareness, the close-set word identification showed a high positive correlation with the level of body-coda awareness and phoneme awareness while the open-set word identification showed a high positive correlation only with the level of phoneme awareness.

• 한국컴퓨터정보학회논문지
• /
• 제15권7호
• /
• pp.83-90
• /
• 2010

어휘 인식 시스템에서 인식률 저하의 요인으로는 유사한 음소 인식과 부정확한 어휘 제공으로 인해 오인식 오류가 존재한다. 부정확한 어휘의 입력으로 특징을 추출하여 인식할 경우 오인식의 결과가 나타나거나 유사한 음소로 인식되며 특징 추출이 제대로 이루어지지 않으면 음소 인식 시 유사한 음소로 인식하게 된다. 따라서 본 논문에서는 음소가 갖는 특징을 기반으로 음소 유사율을 이용한 어휘 인식 후처리에서의 오류 보정 후처리 시스템을 제안하였다. 음소 유사율은 모노폰으로 훈련시킨 훈련 데이터를 각각의 음소에 MFCC와 LPC 특징 추출 방법을 이용하여 구하였다. 유사한 음소는 정확한 음소로 인식할 수 있도록 유도하여 부정확한 어휘 제공으로 인하여 오인식되는 오류를 최소화하였다. 음소 유사율과 신뢰도를 이용하여 오류 보정율을 구하였으며, 어휘 인식 과정에서 오류로 판명된 어휘에 대하여 오류 보정을 수행하였다. 에러패턴 학습을 이용한 시스템과 의미기반을 이용한 시스템에 비해 시스템 성능 평가 결과 MFCC와 LPC는 각각 7.5%와 5.3%의 인식 향상률을 보였다.

• 한국ITS학회 논문지
• /
• 제6권3호
• /
• pp.77-86
• /
• 2007

최근 텔레매틱스 단말기 등과 같이 음성인식을 인터페이스로 하는 음성기반 검색시스템들이 많이 개발되고 있다. 그러나 음성인식에는 여전히 많은 오류가 존재하며, 이에 오류보정에 대한 여러 가지 연구가 진행되고 있다. 본 논문에서는 한국어의 음소가 갖는 특징을 기반으로 음성인식 후처리에서의 오류보정을 제안하였다. 이를 위해 한국어 음소의 특징을 고려한 음소 유사율을 사용하였다. 음소 유사율은 훈련데이터를 모노폰으로 훈련시켜 한국어 음소 각각에 대하여 MFCC와 LPC 특징추출방법을 사용하여 특징추출을 수행하고, 바타차랴 거리 측정법을 사용하여 각 음소 사이의 유사율을 구하였다. 음소 유사율과 신뢰도를 이용하여 오류보정률을 구하였으며, 이를 사용하여 음성인식 과정에서 오류로 판명된 어절에 대하여 오류보정을 수행하고, 음절 복원과 형태소 분석을 재수행하는 과정을 거쳤다. 실험 결과 MFCC와 LPC 각각 7.5%와 5.3%의 인식 향상률을 보였다.

• 한국음향학회:학술대회논문집
• /
• 한국음향학회 1992년도 학술논문발표회 논문집 제11권 1호
• /
• pp.61-65
• /
• 1992

This paper presents a study on Neural Networks for Phoneme Recognition and performs phoneme recognition using TDNN(Time Delay Neural Network). Also, this paper proposes new training algorithm for speech recognition using neural nets that proper to large scale TDNN. Because phoneme recognition is indispensable for continuous speech recognition, this paper uses TDNN to get accurate recognition result of phoneme. And this paper proposes new training algorithm that can converge TDNN to optimal state regardless of the number of phoneme to be recognized. The result of recognition on three phoneme classes shows recognition rate of 9.1%. And this paper proves that proposed algorithm is a efficient method for high performance and reducing convergence time.

• 대한음성학회:학술대회논문집
• /
• 대한음성학회 2006년도 춘계 학술대회 발표논문집
• /
• pp.71-74
• /
• 2006

The purpose of this study is to observe how Korean listeners detect a target phoneme with 'Focus' represented by prosodic prominence and question-induced semantic emphasis. According to the automated phoneme detection task using E-Prime, Korean listeners detected phoneme targets more rapidly when the target-bearing words were in prominence position and in question-induced position. However, when phoneme targets were in prominence position, response time was much faster than in question-induced position. The results suggest that the prosodic prominence which is explicit method of focus representation be more effective than question-inducing, implicit method of it, in phoneme detecting.

• 한국컴퓨터정보학회논문지
• /
• 제21권9호
• /
• pp.27-36
• /
• 2016

Recognition of grapheme is a very important process in the recognition within 'Hangul(Korean written language)' letters using phoneme recognition. It is because the success or failure in the recognition of phoneme greatly affects the recognition of letters. For this reason, it is reported that separation of phonemes is the biggest difficulty in the phoneme recognition study. The current study separates and suggests the new phonemes that used the connective elements that are helpful for dividing phonemes, recommends the features for recognition of such suggested phonemes, databases this, and carried out a set of experiments of recognizing phonemes using the suggested features. The current study used 350 letters in the experiment of phoneme separation and recognition. In this particular kind of letters, there were 1,125 phonemes suggested. In the phoneme separation experiment, the phonemes were divided in the rate of 100%, and the phoneme recognition experiment showed the recognition rate of 98% in recognizing only 14 phonemes into different ones.

• 전자공학회논문지S
• /
• 제34S권4호
• /
• pp.33-41
• /
• 1997

Korean unvoiced phonemes consist of nonstationary parts comparing that the vowels and nasal consonants consist of quasi-stationary part. And some phonemes, which have smae point of articulation but differnt manner of articulation, has similar characteristics, so it makes to be hard to distinguish each other. A new method usin gchanges and characteristics of acoustic properties of these phonemes to improve recognition rate are proposed. And because these changes and cahracteristics evidently occur in continuous speech except some unvoiced consonants are articulated as voiced phoneme in case to be used as an midial between voiced phonemes, this method can be applied easily. The features of the frames extracted to represent each phonemes are used asinputs to the hierarchical neural network. And with these results final decision for phoneme recognition is made thorugh post processing which the new method is applied to. Through the experimental recognition results for 9 unvoiced consonants which belong to bilabial, alveolar, and velar phoneme series, 89.4% recognition rate to distinguish in same phoneme series is obtained, and 85.6% recognition rate is obtained in case of including cistinguishing phoneme series.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2000년도 ITC-CSCC -2
• /
• pp.629-632
• /
• 2000

In this paper, we studied on the phoneme classification for Korean speech recognition. In the case of making large vocabulary speech recognition system, it is better to use phoneme than syllable or word as recognition unit. And, In order to study the difference of speech recognition according to the number of phoneme as recognition unit, we used the speech toolkit of OGI in U.S.A as recognition system. The result showed that the performance of diphthong being unified was better than that of seperated diphthongs, and we required the better result when we used the biphone than when using mono-phone as recognition unit.

• 한국정보처리학회논문지
• /
• 제7권7호
• /
• pp.2246-2253
• /
• 2000

In this paper, we show that one can enhance the performance of blind segmentation of phoneme boundaries by adopting the knowledge of Korean syllabic structure and the regions of voiced/unvoiced sounds. eh proposed method consists of three processes : the process to extract candidate phoneme boundaries, the process to detect boundaries of voiced/unvoiced sounds, and the process to select final phoneme boundaries. The candidate phoneme boudaries are extracted by clustering method based on similarity between two adjacent clusters. The employed similarity measure in this a process is the ratio of the probability density of adjacent clusters. To detect he boundaries of voiced/unvoiced sounds, we first compute the power density spectrum of speech signal in 0∼400 Hz frequency band. Then the points where this paper density spectrum variation is greater than the threshold are chosen as the boundaries of voiced/unvoiced sounds. The final phoneme boundaries consist of all the candidate phoneme boundaries in voiced region and limited number of candidate phoneme boundaries in unvoiced region. The experimental result showed about 40% decrease of insertion rate compared to the blind segmentation method we adopted.