한국정보기술학회
한국정보기술학회논문지 제9권 제2호
2011.2, 43-55 (총13페이지)
Reliability-weighted HMM Considering Inexact Observation For Enhancing Recognition Performance
Md. Tariquzzaman, Jin Young Kim, Seungyou Na, Hyoung-Gook Kim, Min Gyu Song
영어 초록
HMM (Hidden Markov Model) is widely used in pattern recognition areas such as speech and speaker recognition, handwritten recognition, gesture recognition, and so on. In this paper, we present a reliability-weighted HMM (RW-HMM) approach considering inexact observations. We introduce a weighting factor - confidence measures of observations - in HMM target function and drive a training algorithm based on the traditional EM (Expectation Maximization) algorithm for optimizing a modified HMM target function. To verify the usefulness of our proposed method, we performed SI (speaker identification) experiments using ETRI speaker recognition database. With the proposed RW-HMM, the experimental results show that the performance of SI is highly enhanced particularly in noisy environments. Finally, RW-HMM could be applied in any applications with well-defined reliability.
한국어 초록
은닉 마코프 모델(HMM)은 음성인식, 화자인식, 필기체 인식, 제스처 인식 등과 같은 패턴 인식 분야에서 널리 사용되는 기술이다. 본 논문에서는 일반적인 HMM을 변형한 부정확한 관측 값들을 고려한 신뢰성 있는 가중 은닉 마코프 모델(RW-HMM)을 제안한다. 우리는 은닉 마코프 모델의 목적 함수에 관측 값들의 신뢰성 척도인 가중 인자를 도입하고, 변형된 HMM의 목적함수를 최적화하기 위해 일반적인 EM알고리듬에 기반 한 학습 알고리듬을 제안한다. 제안된 방법의 유용성을 검증하기 위해 본 논문에서는 ETRI의 화자 인식 데이터베이스를 이용하여 화자 인증 실험을 수행하였다. 실험 결과, 제안된 알고리듬이 기존의 방법에 비해 성능 향상을 보였고, 특히 잡음 환경에서 더 높은 성능 향상을 보였다. 결론적으로 제안된 RW-HMM은 잘 정의된 신뢰성을 바탕으로 어떠한 환경에서도 적용될 수 있을 것으로 생각된다.
목차
Abstract
요약
Ⅰ. Introduction
Ⅱ. Reliability-Weighted Hidden Markov Model(RW-HMM)
Ⅲ. Validation Experiments of RW-HMM in Speaker Identification Domain
Ⅳ. Conclusion
References
Authors
키워드
observation reliability, hidden markov model, EM algorithm, speaker identification

참고문헌 (19)

  • J. P. Campbell , 1997 , Speaker recognition: a tutorial , Proceedings of the IEEE 85 (9) : 1437 ~ 1462
  • D. A. Reynolds , 2000 , An overviews of automatic speaker recognition technology , ICASSP 4 : 4072 ~ 4075
  • B. Zhen , 2000 , An enhanced RASTA processing for speaker identification , ICSLP : 251 ~ 254
  • R. J. Mammone , 1996 , Robust speaker recognition, a feature-based approach , IEEE Signal Processing Magazine 13 (5) : 58 ~ 71
  • D. Stephane , 2003 , Robust feature extraction and acoustic modeling at multitel: experiments on the Auroradatabases , Eurospeech : 1789 ~ 1792
  • A. Rosenberg , 1994 , Cepstral channel normalization techniques for HMM-based speaker verification , ICSLP : 1835 ~ 1838
  • E. Mengusoglu , 2003 , Confidence measure based model adaptation for speaker verification , 2nd IASTED International Conference on Communications, Internet and Information Technology : 408 ~ 411
  • Chin-Hung Sit , 2004 , Maximum likelihood and maximum a posteriori adaptation for distributed speaker recognition systems , 1st Int. Conf. on Biometric Authentication
  • K. Yiu , 2003 , Environment adaptation for robust speaker verification , EUROSPEECH- 2003 : 2973 ~ 2976
  • R. Chengalvarayan , 2001 , A maximum a posteriori approach to speaker adaptation using the trended hidden Markov model , IEEE Transaction on Speech and Audio Processing 9 (6) : 549 ~ 557
  • C.H. Lee , 1991 , A Study on speaker adatation on the parameters of continuous density hidden Markov models , IEEE Transaction of Signal Processing 39 : 806 ~ 814
  • D.Ramos-Castro,et.al , 2007 , Speaker Verification Using Speaker- and Test- Dependent Fast Score Normalization , Pattern Recognition Letters 28 : 90 ~ 98
  • R. Auckenthaler , 2000 , Score Normalization for Text-Independent Speaker Verification Systems , Digital Signal Processing 10 : 42 ~ 54
  • Jin Young Kim, So Hee Min, Seung You Na , 2007 , Modified GMM Training for Inexact Observation and Its Application to Speaker Identification , 음성과학 14 (1) : 163 ~ 175
  • R. Rabiner , 1986 , An introduction to hidden Markov models , ASSP Magazine 3 (1) : 4 ~ 16
  • M. Brand , 1997 , Coupled hidden Markov models for complex action reconition , IEEE International Conference on Computer Vision and Pattern Recogntion : 994 ~ 999
  • Z. Ghahramani , 1997 , Factorial hidden Markov models , Machine Learning 29 (2) : 245 ~ 273
  • L. A. Liporace , 1982 , Maximum likelihood estimation for multivariate observations of Markov sources , IEEE Trans. On Information Theory IT-28 (5) : 729 ~ 734
  • R. Eberhart , 1995 , A new optimizer using particle swarm theory , Sixth International Symposium on Micro Machine and Human Science : 39 ~ 43

이 논문이 수록된 권·호 내 논문

한국정보기술학회논문지 제9권 제2호