가산잡음환경에서 강인음성인식을 위한 은닉 마르코프 모델 기반 손실 특징 복원
HMM-based missing feature reconstruction for robust speech recognition in additive noise environments
2
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
This paper describes a robust speech recognition technique by reconstructing spectral components mismatched with atraining environment. Although the cluster-based reconstruction method can compensate the unreliable components fromreliable components in the same spectral vector by assuming an independent, identically distributed Gaussian-mixture processof training spectral vectors, the presented method exploits the temporal dependency of speech to reconstruct the componentsby introducing a hidden-Markov-model prior which incorporates an internal state transition plausible for an observed spectralvector sequence. The experimental results indicate that the described method can provide temporally consistent reconstructionand further improve recognition performance on average compared to the conventional method.
Abstract
This paper describes a robust speech recognition technique by reconstructing spectral components mismatched with atraining environment. Although the cluster-based reconstruction method can compensate the unreliable components fromreliable components in the same spectral vector by assuming an independent, identically distributed Gaussian-mixture processof training spectral vectors, the presented method exploits the temporal dependency of speech to reconstruct the componentsby introducing a hidden-Markov-model prior which incorporates an internal state transition plausible for an observed spectralvector sequence. The experimental results indicate that the described method can provide temporally consistent reconstructionand further improve recognition performance on average compared to the conventional method.
참고문헌 (11)
- Acero, A., "Acoustic and Environmental Robustness in Automatic Speech Recognition", PhD, Dept. of Electrical and Computer Engineering, pp.1990
- Raj, B., "Missing feature approaches in speech recognition", IEEE Signal Processing Magazine, 22, pp.101~116 (2005)
- Peinado, A. M., "MMSE-based Channel Mitigation for Distributed Speech Recognition", Proc. EUROSPEECH, pp.2707~2710 (2001)
- Peinado, A. M., "HMM-Based Methods for Channel Error Mitigation in Distributed Speech Recognition", Proc. ICSLP02, pp.2205~2208 (2002)
- Borgström, B. J., "HMM-based reconstruction of unreliable spectrographic data for noise robust speech recognition", IEEE Transactions on Audio, Speech, and Language Processing, 18, pp.1612~1623 (2010)
- Huang, X., "Spoken language processing : a guide to theory, algorithm, and system development" (2001), Prentice-Hall
- Cho, J. -W., "An efficient HMM-based feature enhancement method with filter esimation for reverberant speech recognition", IEEE Signal Processing Letter, 20, pp.1199~1202 (2013)
- Price, P., "The DARPA 1000-Word Resource Management Database for Continuous Speech Recognition", Proc. IEEE ICASSP, pp.651~654 (1988)
- Young, S., "The HTK book" (2006), Cambridge University Press
- Varga, A., "Assessment for automatic speech recognition: 2. In: NOISEX 1992: A Database and anExpeiment to Study the Effect of Additive Noise on Speech Recognition Systems", Speech Comm, 12, pp.247~251 (1993)