• 한국음향학회지
• /
• 제29권2호
• /
• pp.148-153
• /
• 2010

음성 신호처리 환경에서 잡음이 섞인 신호를 개선할 목적으로 음성향상 기법이 많이 이용되고 있다. 잡음추정 알고리즘은 변화하는 환경에 빠르게 적응할 수 있어야 하며 음성신호의 영향을 줄이기 위해 음성신호가 존재하지 않는 구간에서만 잡음의 파워를 갱신한다. 이러한 방법은 음성구간검출이 선행되어야 한다. 그러나 잡음에 열화된 음성신호에 묵음구간이 존재하지 않을 경우, 위와 같이 음성검출을 통한 묵음구간에서의 잡음 추정 방법 및 SNR 추정 방법이 적용될 수 없다. 본 논문에서는 묵읍구간이 존재하지 않는 연속음성신호에서 SNR을 추정하는 기법을 제안한다. 음성신호는 MBE(Multi-Band Excitation) 발성 모델에 따라 유 무성음으로 구분할 수 있다. 그리고 에너지가 유성음에 대부분 분포하기 때문에, 부가성 잡음환경에서 유성음의 에너지를 음성신호의 에너지로 근사화하여 SNR을 추정할 수 있다. 제안하는 방식은 연속음성신호를 IMBE (Improved Multi-Band Exciation) 보코더를 이용해 유 무성음 대역으로 구분하고, 각각 대역의 에너지 정보를 아용하여 단구간 음성신호의 SNR을 계산한다. 전체 음성구간의 SNR은 단구간 SNR의 평균값을 통해 추정한다.

• 대한음성학회지:말소리
• /
• 제64호
• /
• pp.137-153
• /
• 2007

The pronunciation lexicon of a continuous speech recognition system should contain enough pronunciation variations to be used for building a search space large enough to contain a correct path, whereas the size of the pronunciation lexicon needs to be constrained for effective decoding and lower perplexities. This paper describes a procedure for selecting pronunciation variations to be included in the lexicon based on the frequencies of the corresponding phonetic rules observed in the training corpus. Likelihood of a phonetic rule's application is estimated using the observation frequency of the rule and is used to control the construction of a pronunciation lexicon. Experiments with various pronunciation lexica show that the proposed method is helpful to improve the speech recognition performance.

• 융합신호처리학회 학술대회논문집
• /
• 한국신호처리시스템학회 2003년도 하계학술대회 논문집
• /
• pp.18-21
• /
• 2003

In general it has been considered to be the difficult problem that we divide continuous speech into short interval with having identical phoneme quality. In this paper we used Gaussian Mixture Model (GMM) related to probability density to divide speech into phonemes, an initial, medial, and final sound. From them we peformed continuous speech recognition. Decision boundary of phonemes is determined by algorithm with maximum frequency in a short interval. Recognition process is performed by Continuous Hidden Markov Model(CHMM), and we compared it with another phoneme divided by eye-measurement. For the experiments result we confirmed that the method we presented is relatively superior in auto-segmentation in korean speech.

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

In this paper, we propose a language modeling approach to improve the performance of a large vocabulary continuous speech recognition system. The proposed approach is based on the active learning framework that helps to select a text corpus from a plenty amount of text data required for language modeling. The perplexity is used as a measure for the corpus selection in the active learning. From the recognition experiments on the task of continuous Korean speech, the speech recognition system employing the language model by the proposed language modeling approach reduces the word error rate by about 6.6 % with less computational complexity than that using a language model constructed with randomly selected texts.

• The Journal of the Acoustical Society of Korea
• /
• 제15권2E호
• /
• pp.14-23
• /
• 1996

This paper is a on the neural network to recognize the phonemes, the weight initialization to reduce learning speed, and LR parser for continuous speech recognition. The neural network spots the phonemes in continuous speech and LR parser parses the output of neural network. The whole phonemes recognized in neural network are divided into several groups which are grouped by the similarity of phonemes, and then each group consists of neural network. Each group of neural network to recognize the phonemes consisits of that recognize the phonemes of their own group and VGNN(Verify Group Neural Network) which judges whether the inputs are their own group or not. The weights of neural network are not initialized with random values but initialized from learning data to reduce learning speed. The LR parsing method applied to this paper is not a method which traces a unique path, but one which traces several possible paths because the output of neural network is not accurate. The parser processes the continuous speech frame by frame as accumulating the output of neural network through several possible paths. If this accumulated path-value drops below the threshold value, this path is deleted in possible parsing paths. This paper applies the continuous speech recognition system to the threshold value, this path is deleted in possible parsing paths. This paper applies the continuous speech recognition system to the continuous Korea digits recognition. The recognition rate of isolated digits is 97% in speaker dependent, and 75% in speaker dependent. The recognition rate of continuous digits is 74% in spaker dependent.

• 음성과학
• /
• 제3권
• /
• pp.119-131
• /
• 1998

We have developed a keyword spotting system for automatic speech editing. This system recognizes the only keyword 'MBC news' and then sends the time information to the host system. We adopted a vocabulary dependent model based on continuous hidden Markov model, and the Viterbi search was used for recognizing the keyword. In recognizing the keyword, the system uses a parallel network where HMM models are connected independently and back-tracking information for reducing false alarms and missing. We especially focused on implementing a stable and practical real-time system.

• 한국음향학회지
• /
• 제13권1호
• /
• pp.24-31
• /
• 1994

본 논문에서는 연속분포 hidden Markov모델을 이용한 화자독립 연속 음성 인식 시스템에 관해 기술한다. 연속분포 모델은 평균과 분산 벡터로 구성되며 음성신호를 직접 모델링하여 양자화 왜곡이 없어진다. 특징벡터는 filter bank 계수 및 그 1, 2차 미분계수를 사용하여 음성신호의 동적 특성을 반영하였다. Segmental K-means 알고리즘을 이용하여 학습하였으며, 연속어 인식에서 가장 문제가 되는 조음화 현상으로 인한 인식률 저하를 막기 위해 앞뒤의 음소를 고려해주는 triphone을 인식단위로 사용하였다. Search 알고리즘으로는 시간 면에서 효율이 좋은 one-pass search 알고리즘을 사용하였다 성능 평가를 위한 회자 독립인식 실험에서 문법이 없을 경우 $83\%$, finite state network을 적용한 경우에는 $94\%$의 인식률을 나타내었다.

• 대한음성학회지:말소리
• /
• 제49호
• /
• pp.107-121
• /
• 2004

In this paper, we describe a cross-morpheme pronunciation variation model which is especially useful for constructing morpheme-based pronunciation lexicon to improve the performance of a Korean LVCSR. There are a lot of pronunciation variations occurring at morpheme boundaries in continuous speech. Since phonemic context together with morphological category and morpheme boundary information affect Korean pronunciation variations, we have distinguished phonological rules that can be applied to phonemes in within-morpheme and cross-morpheme. The results of 33K-morpheme Korean CSR experiments show that an absolute reduction of 1.45% in WER from the baseline performance of 18.42% WER was achieved by modeling proposed pronunciation variations with a possible multiple context-dependent pronunciation lexicon.

• 전자공학회논문지B
• /
• 제28B권7호
• /
• pp.548-556
• /
• 1991

In this paper, Korean spoken continuous digits are ercognized using DHMM(Dynamic Hidden Markov Model) and lexical analysis to provide the base of developing voice dialing system. After segmentation by phoneme unit, it is recognized. This system can be divided into the segmentation section, the design of standard speech section, the recognition section, and the lexical analysis section. In the segmentation section, it is segmented using the ZCR, O order LPC cepstrum, and Ai, parameter of voice speech dectaction, which is changed according to time. In the standard speech design section, 19 phonemes or syllables are trained by DHMM and designed as a standard speech. In the recognition section, phomeme stream are recognized by the Viterbi algorithm.In the lexical decoder section, finally recognized continuous digits are outputed. This experiment shiwed the recognition rate of 85.1% using data spoken 7 times of 21 classes of 7 continuous digits which are combinated all of the occurence, spoken by 10 man.

• 대한전자공학회논문지SP
• /
• 제41권3호
• /
• pp.201-207
• /
• 2004

본 논문은 연속음 인식에 관한 연구로써, 구문 분석을 이용한 One-Stage DMS/DP에 의한 음성 인식 방법을 사용한다. 인식 실험을 위해 우선 구간 구분화 알고리즘을 이용하여 DMS(dynamic Multi-Section) 모델을 만들며, 구문 분석을 이용한 One-Stage DMS/DP 방법으로 연속음 데이터를 인식하게 하였다. 제안된 방법에 의한 인식 실험을 수행하는 것 외에도 비교를 위해 전통적인 One-Stage DP 방법을 같은 조건 같은 데이터를 가지고 수행하였다. 인식 실험 결과, 기존의 방법보다 구문분석을 이용한 One-Stage DMS/DP 방법이 우수한 것으로 나타났다.