• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07b
• /
• pp.853-855
• /
• 2005

은닉 마르코프(HMM)의 HMM의 구조 최적을 위한 모델 선택 방법에 많은 방법들이 연구되어지고 있다. HMM의 구조를 어떻게 최적으로 정해야 하는 가에 대해 HMM의 구조를 체계적인 방법으로 정함과 동시에 변별력의 단점을 개선 할 수 있는 방법으로 Anti-likelihood(ALC1)를 제안하였고 이를 모델 선택 기준인 BIC와의 결합(ALC2)하여 필기 데이터에 대해 실험한 결과 기존의 방법보다 파라미터의 수는 감소되고 인식률이 향상됨을 알 수 있었다. 이를 Discrete HMM에도 적용하여 제안된 ALC2가 HMM 구조를 최적화하는 모델 선택 기준임을 Continuous-Density HMM과 비교하여 실험 검증 한다.

• The Journal of the Acoustical Society of Korea
• /
• v.26 no.4
• /
• pp.144-152
• /
• 2007

Conventional research works on the classification of the heart sound signal have been done mainly with the artificial neural networks. But the analysis results on the statistical characteristic of the heart sound signal have shown that the HMM is suitable for modeling the heart sound signal. In this paper, we model the various heart sound signals representing different heart diseases with the HMM and find that the classification rate is much affected by the clustering of the heart sound signal. Also, the heart sound signal acquired in real environments is a continuous signal without any specified starting and ending points of time. Hence, for the classification based on the HMM, the continuous cyclic heart sound signal needs to be manually segmented to obtain isolated cycles of the signal. As the manual segmentation will incur the errors in the segmentation and will not be adequate for real time processing, we propose a variant of the ergodic HMM which does not need segmentation procedures. Simulation results show that the proposed method successfully classifies continuous heart sounds with high accuracy.

• The Journal of the Acoustical Society of Korea
• /
• v.15 no.4
• /
• pp.5-11
• /
• 1996

In this study, the method of speaker adaptation for uttered sentence using syllable unit hmm is proposed. Segmentation of syllable unit for sentence is performed automatically by concatenation of syllable unit hmm and viterbi segmentation. Speaker adaptation is performed using MAPE(Maximum A Posteriori Probabillity Estimation) which can adapt any small amount of adaptation speech data and add one sequentially. For newspaper editorial continuous speech, the recognition rates of adaptation of HMM was 71.8% which is approximately 37% improvement over that of unadapted HMM

• Journal of the Institute of Electronics Engineers of Korea TE
• /
• v.37 no.3
• /
• pp.16-20
• /
• 2000

Most of speech recognition systems are using Hidden Markov Model based on statistical modelling frequently. In Korean isolated telephone digit speech recognition, high recognition rate is gained by using HMM if many training data are given. But in Korean continuous telephone digit speech recognition, HMM has some limitations for similar telephone digits. In this paper we suggest a way to overcome some limitations of HMM by using discriminant learning based on minimal classification error criterion in Korean continuous telephone digit speech recognition. The experimental results show our method has high recognition rate for similar telephone digits.

• The Journal of the Acoustical Society of Korea
• /
• v.14 no.1
• /
• pp.81-89
• /
• 1995

In this paper, we report the continuous speech recognition system using the continuous HMM with discrete duration control and the regression coefficients. Also, we do recognition experiment using One Pass DP method(for 25 sentences of robot control commands) with finite state automata context control. In the experiment for 4 connected spoken digits, the recognition rates are $93.8\%$ when the discrete duration control and the regression coefficients are included, and $80.7\%$ when they are not included. In the experiment for 25 sentences of the robot control commands, the recognition rate are $90.9\%$ when FSN is not included and $98.4\%$ when FSN is included.

• The Journal of the Acoustical Society of Korea
• /
• v.15 no.2
• /
• pp.85-89
• /
• 1996

In this paper, transition constrained Hidden Markov Model(HMM) in which the transition between states occur only within prescribed time slot is proposed and the performance is evaluated in the noisy environment. The transition constrained HMM can explicitly limit the state durations and accurately de scribe the temporal structure of speech signal simply and efficiently. The transition constrained HMM is not only superior to the conventional HMM but also require much less computation time. In order to evaluate the performance of the transition constrained HMM, speaker independent isolated word recognition experiments were conducted using semi-continuous HMM with the noisy speech for 20, 10, 0 dB SNR. Experiment results show that the proposed method is robust to the environmental noise. The 81.08% and 75.36% word recognition rates for conventional HMM was increased by 7.31% and 10.35%, respectively, by using transition constrained HMM when two kinds of noises are added with 10dB SNR.

• Speech Sciences
• /
• v.7 no.4
• /
• pp.11-26
• /
• 2000

In this paper, we propose the HMM with the MIN module. Because initial and re-estimated variance vectors are important elements for performance in HMM recognition systems, we propose a method which compensates for the mismatched statistical feature of training and test data. The MIN module function is a differentiable function similar to the sigmoid function. Unlike a continuous density function, it does not include variance vectors of the data set. The proposed hybrid HMM/MIN module is a unified network in which the observation probability in the HMM is replaced by the MIN module neural network. The parameters in the unified network are re-estimated by the gradient descent method for the Maximum Likelihood (ML) criterion. In estimating parameters, the variance vector is not estimated because there is no variance element in the MIN module function. The experiment was performed to compare the performance of the proposed HMM and the conventional HMM. The experiment measured an isolated number for speaker independent recognition.

• The Journal of the Acoustical Society of Korea
• /
• v.11 no.6
• /
• pp.46-52
• /
• 1992

본 논문에서는 HMM 의 이산분포를 연속분포로 근사시키는 준 연속분포 HMM 에 의한 한국어 단어인식에 관하여 연구하였다. 이 모델의 생성과정에서는 입력벡터의 출력확률을 혼합 다차원 정규분 포로 가정하여 입력벡터의 확률함수와 코드위드의 심볼출력을 선형결합하므로써, 연속분포 모델로 근사 시켰으며, 단어인식과정에서는 생성모델에 의해 이산분포 모델에서 발생되는 양자와 왜곡을 감소시키므 로써 인식률을 향상시켰다. 이 방법을 평가하기 위하여 DDD 지역명을 대상으로 이산분포 HMM과 준연 속분포 HMM 의 비교실험을 수행하였다. 그 결과 준연속분포 HMM 에 의하여 이산분포 HMM 보다 향상된 인식률을 얻을 수 있었다.

• Journal of IKEEE
• /
• v.11 no.2
• /
• pp.89-94
• /
• 2007

This paper is a study on continuous speech recognition in the Korean language using HMM-based models with continuous density functions. Here, we propose the most efficient method of continuous speech recognition for the Korean language under the condition of a continuous HMM model with 2 to 44 density functions. Two voice models were used CI-Model that uses 36 uni-phones and CD-Model that uses 3,000 tri-phones. Language model was based on N-gram. Using these models, 500 sentences and 6,486 words under speaker-independent condition were processed. In the case of the CI-Model, the maximum word recognition rate was 94.4% and sentence recognition rate was 64.6%. For the CD-Model, word recognition rate was 98.2% and sentence recognition rate was 73.6%. The recognition rate of CD-Model we obtained was stable.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1994.06c
• /
• pp.157-160
• /
• 1994

본 논문에서는 한국어 숫자음 인식에 대한 이산분포 HMM과 연속분포 HMM의 인식 성능을 비교하였다. 일반적으로 연속분포 HMM은 많은 계산량이 필요하고, 학습시 초기값이 매우 민감하다는 단점이 있지만, 이산분포 HMM의 VQ로 인한 왜곡을 제거함으로써 인식률을 향상시킬 수 있다. 여기서는 성능비교를 위해서 mel-cepstrum의 분석차수, 이산분포 HMM의 codebook 크기, 연속분포 HMM의 miture 개수등에 따른 인식성능을 비교하였다. 실험 결과 이산분포 HMM에서는 mel-cepstrum 벡터가 14차이고, codebook 크기가 64일 때 가장 좋은 성능을 나타냈으며, 연속부포 HMM에서는 mel-cepstrum 벡터가 16차이고 miture가 3개일 때 가장 좋은 결과를 얻을 수 있었다. 특히 학습 데이터의 양이 적은 경우에는 연속분포 HMM이 이산분포 HMM보다 더 좋은 인식률을 나타내었다.