• 전자공학회논문지B
• /
• 제30B권2호
• /
• pp.51-58
• /
• 1993

This paper describes the development of a speaker-dependent isolated word recognizer as applied to voice dialing in a car noise environment. for this purpose, several methods to improve performance under such condition are evaluated using database collected in a small car moving at 100km/h The main features of the recognizer are as follow: The endpoint detection error can be reduced by using the magnitude of the signal which is inverse filtered by the AR model of the background noise, and it can be compensated by using variants of the DTW algorithm. To remove the noise, an autocorrelation subtraction method is used with the constraint that residual energy obtainable by linear predictive analysis should be positive. By using the noise rubust distance measure, distortion of the feature vector is minimized. The speech recognizer is implemented using the Motorola DSP56001(24-bit general purpose digital signal processor). The recognition database is composed of 50 Korean names spoken by 3 male speakers. The recognition error rate of the system is reduced to 4.3% using a single reference pattern for each word and 1.5% using 2 reference patterns for each word.

• 대한음성학회:학술대회논문집
• /
• 대한음성학회 2005년도 춘계 학술대회 발표논문집
• /
• pp.49-52
• /
• 2005

Cepstral Mean Subtraction (CMS) makes effectively compensation for a channel distortion, but there are some shortcomings such as distortions of feature parameters, waiting for the whole speech sentence. By assuming that the silence parts have the channel characteristics, we consider the channel normalization using subtraction of cepstral means which are only obtained in the silence areas. If the considered techniques are successfully used for the channel compensation, the proposed method can be used for real time processing environments or time important areas. In the experiment result, however, the performance of our method is not good as CMS technique. From the analysis of the results, we found potentiality of the proposed method and will try to find the technique reducing the gap between CMS and ours method.

• 음성과학
• /
• 제10권3호
• /
• pp.133-144
• /
• 2003

The purpose of this acoustic experimental study was to investigate interrelation between prosodic features of Korean spoken language and those of Korean folk songs. The words of Changbutaryoung were spoken for analysis of spoken language by three female graduate students and the song was sung for musical features by three Kyunggi Minyo singers. Pitch contours were analyzed from sound spectrogram made by Pitch Works. Results showed that special musical voices (breaking, tinkling, vibrating, etc.) and tunes (rising, falling, level, etc) of folk song were discovered at the same place where accents of spoken language came. It appeared that, even though the patterns of pitch contour were different from each other, there was positive interrelation between prosodic features of Korean spoken language and those of Korean folk songs.

• 말소리와 음성과학
• /
• 제3권3호
• /
• pp.57-62
• /
• 2011

Focused items in English sentences are usually accompanied by changes in acoustic manifestation. This paper investigates the acoustic characteristics of $H^*$ in English utterances produced by natives speakers of English and Korean learners of English. To obtain more reliable results, the changes of the acoustic feature values (F0, intensity, syllable duration) were normalized by a median value and a whole duration of each utterance. Acoustic values of sentences with no focused words were compared with those of sentences with focused words within each group (Americans vs. Koreans). Sentences with focused words were compared between the two groups, too. In the instances in which a significant Group x Focus Location (initial, middle and final of a sentence) interaction was obtained, further analysis testing the effect of Group on each Focus Location was conducted. The analysis revealed that Korean learners of English produced focused words with lower F0, lower intensity and shorter syllable duration than native speakers of English. However, the effect of intensity change caused by focus was not significant within each group. Further analysis examining the interaction of Group and Focus Location showed that the change in F0 produced by Korean group was significantly lower in the middle and the final positions of sentences than by American group. Implications for the intonation training were also discussed.

• 대한음성학회:학술대회논문집
• /
• 대한음성학회 2005년도 춘계 학술대회 발표논문집
• /
• pp.103-106
• /
• 2005

The goal of our research is to build a text independent speaker recognition system that can be used in any condition without any additional adaptation process. The performance of speaker recognition systems can be severally degraded in some unknown mismatched microphone and noise conditions. In this paper, we show that PCA(Principal component analysis) without dimension reduction can greatly increase the performance to a level close to matched condition. The error rate is reduced more by the proposed augmented PCA, which augment an axis to the feature vectors of the most confusable pairs of speakers before PCA

• 한국음향학회지
• /
• 제3권1호
• /
• pp.25-35
• /
• 1984

This paper analysis the Dynamic Time Warping algorithms for time normalization of speech pattern and discusses the Dynamic Programming algorithm for spoken Korean isolated digits recognition. In the DP matching, feature vectors of the reference and test pattern are consisted of first three formant frequencies extracted by power spectrum density estimation algorithm of the ARMA model. The major differences in the various DTW algorithms include the global path constrains, the local continuity constraints on the path, and the distance weighting/normalization used to give the overall minimum distance. The performance criterias to evaluate these DP algorithms are memory requirement, speed of implementation, and recognition accuracy.

• 대한구순구개열학회지
• /
• 제5권2호
• /
• pp.95-108
• /
• 2002

언어 발달의 조기 단계를 이해하기 위한 일환으로 crying은 언어전 발달의 기초 단계로서 여러 학문적 분야에서 많은 연구가 있어왔다. 그러나 구순구개열(CLP))환아의 경우는cry-producing/control mechnism에 variation이 많은 이유로 이 분야의 연구는 거의 없는 실정이다. 이에 본 연구에서는 다음과 같은 의문점을 가지고 CLP환아의 cry feature에 대한분석을 하였다. 첫째, 정상아와 CLP환아의 cry에 전형적인 차이가 있는가? 둘째, CLP환아의 술전, 술후 cry feature에 변화가 있는가? 셋째, cry분석이 CLP환아의 이후 speech disorder에 대한 언어전 평가로서의 가치가 있는가? 넷째, 특정 parameter가 언어전 평가에 적절한 도구로 작용할 수 있는가? 생후 15개월 이내의 CLP 환아 3명과 유사한 나이대의 정상아 8명의 cry에 대한 공기역학 및 음향음성학적 분석을 통해 CLP 환아와 정상아, CLP환아의 술전, 술후 cry특성을 비교 분석하였다. 결과는 다음과 같다. 1 공기역학적 분석 1) airflow는 CLP 환아의 경우 정상아보다 약간 높았고 술 후 약간 증가하였다. 2)폐활량을 나타내는volume에서는 정상아보다 술전 CLP환자의 경우 보상적으로 더 큰 수치를 보였고 술후 약간 증가하였다. 3)강도를 나타내는 parameter(SPL)에서는 정상아 보다 술전 CLP환자의 계측치가 약간 작았으나 술 후 증가하는 양상을 보였다. 2. 음향음성학적 분석 1)기저 주파수 분석시 정상아에 비해 술 전 CLP환자의 경우 계측치가 약간 낮았으나 술 후 증가하여 정상군의 계측치에 근접하였다. 2)강도를 나타내는energy 측정시 정상아에 비해 술 전 CLP계측치가 보상성으로 약간 큰수치를 나타내었고 술 후 약간 더 증가하였다. 3) Shimmer에서는CUI환자의 술후계측치가술전에 비해 현저히 감소하여 정상군의 수치에 근접하였다.

• 한국지능시스템학회논문지
• /
• 제18권1호
• /
• pp.20-26
• /
• 2008

지능형 로봇이나 컴퓨터가 일상생활 속에서 차지하는 비중이 점점 높아짐에 따라 인간과의 상호교류도 점점 중요시되고 있다. 이렇게 지능형 로봇(컴퓨터) - 인간의 상호 교류하는데 있어서 감정 인식 및 표현은 필수라 할 수 있겠다. 본 논문에서는 음성 신호와 얼굴 영상에서 감정적인 특징들을 추출한 후 이것을 Bayesian Learning과 Principal Component Analysis에 적용하여 5가지 감정(평활, 기쁨, 슬픔, 화남, 놀람)으로 패턴을 분류하였다. 그리고 각각 매개체의 단점을 보완하고 인식률을 높이기 위해서 결정 융합 방법과 특징 융합 방법을 적용하여 감정 인식 실험을 하였다. 결정 융합 방법은 각각 인식 시스템을 통해 얻어진 인식 결과 값을 퍼지 소속 함수에 적용하여 감정 인식 실험을 하였으며, 특징 융합 방법은 SFS(Sequential Forward Selection) 특징 선택 방법을 통해 우수한 특징들을 선택한 후 MLP(Multi Layer Perceptron) 기반 신경망(Neural Networks)에 적용하여 감정 인식 실험을 실행하였다. 그리고 인식된 결과 값을 2D 얼굴 형태에 적용하여 감정을 표현하였다.

• 한국정보통신학회논문지
• /
• 제9권8호
• /
• pp.1625-1631
• /
• 2005

본 논문에서는 언어모델과 음향모델을 개선함으로써 단위 숫자음의 인식성능 최적화에 대해 설명한다. 언어모델은 한국어 단위 숫자음 문장의 문법적 특징을 분석하고, FSN 노드를 두음절로 구성하여 오 인식률을 감소시켰다. 음향모델은 단음절로 구성되어 발성기간이 짧고 조음이 많이 생기는 불명확한 음소, 음절의 분할로 연한 오 인식을 줄이기 위해 인식단위를 반음절쌍으로 하였다. 인식단위의 특징을 효과적으로 모델링하기 위해 특징레벨에서 K-means 알고리즘(4)으로 클러스터링 하여 상태를 분할하는 변형된 연쇄 상태 분할방법을 이용하였다. 실험 결과 제안된 언어모델의 적용 후 동일 문백종속 음소모델에서 $10.5\%$, 음향모델에서 인식단위를 반음절쌍으로 하였을 경우 문백종속 음소모델에 비해 $12.5\%$, 변형된 연쇄 상태분할을 하였을 경우 $1.5\%$의 인식률을 향상시킬 수 있었다.

• The Journal of the Acoustical Society of Korea
• /
• 제25권1E호
• /
• pp.20-24
• /
• 2006

In this paper, we focused on the real time implementation of a speech recognition system with medium size of vocabulary considering its application to a mobile phone. First, we developed the PC based variable vocabulary word recognizer having the size of program memory and total acoustic models as small as possible. To reduce the memory size of acoustic models, linear discriminant analysis and phonetic tied mixture were applied in the feature selection process and training HMMs, respectively. In addition, state based Gaussian selection method with the real time cepstral normalization was used for reduction of computational load and robust recognition. Then, we verified the real-time operation of the implemented recognition system on the TMS320C6201 EVM board. The implemented recognition system uses memory size of about 610 kbytes including both program memory and data memory. The recognition rate was 95.86% for ETRI 445DB, and 96.4%, 97.92%, 87.04% for three kinds of name databases collected through the mobile phones.