Signal Trajectory Based Noise Compensation for Robust Speech Recognition
This paper presents a novel signal trajectory based noise compensation algorithm for robust speech recognition. Its performance is evaluated on the Aurora 2 database. The algorithm consists of two processing stages: 1) noise spectrum is estimated using trajectory auto-segmentation and clustering, so that spectral subtraction can be performed to roughly estimate the clean speech trajectories; 2) these trajectories are regenerated using trajectory HMMs, where the constraint between static and dynamic spectral information is imposed to refine the noise subtracted trajectories both in “level” and “shape”. Experimental results show that the recognition performance after spectral subtraction is improved with or without trajectory regeneration, but the HMM regenerated trajectories yields the best performance improvement. After spectral subtraction, the average relative error rate reductions of clean and multi-condition training are 23.21% and 5.58%, respectively. And the proposed trajectory regeneration algorithm further improves them to 42.59% and 15.80%.
KeywordsTrajectory Regeneration Clean Speech Noisy Speech Voice Activity Detector Speech Recognizer
Unable to display preview. Download preview PDF.
- 2.Ealey, D., Kelleher, H., Pearce, D.: Harmonic Tunneling: Tracking Non-Stationary Noises During Speech. In: Proc. Eurospeech 2001, Scandinavia, pp. 437–440 (2001)Google Scholar
- 3.Evans, N.W.D., Mason, J.S.: Time-Frequency Quantile-Based Noise Estimation. In: Proc. EUSIPCO, Toulouse, vol. 1, pp. 539–542 (2002)Google Scholar
- 4.Yang, C., Soong, F.K., Lee, T.: Static and Dynamic Spectral Features: Their Noise Robustness and Optimal Weights for ASR. In: Proc. ICASSP 2005, Philadelphia, vol. 1, pp. 241–244 (2005)Google Scholar
- 5.Tokuda, K., Yoshimura, T., Masuko, T., Kobayashi, T., Kitamura, T.: Speech parameter generation algorithms for HMM-based speech synthesis. In: Proc. ICASSP 2000, Istanbul, vol. 3, pp. 1315–1318 (2000)Google Scholar
- 6.Young, S.J., Evermann, G., et al.: The HTK Book, Revised for HTK Version 3.3 (2005)Google Scholar
- 7.Hirsch, H.G., Pearce, D.: The Aurora Experimental Framework for the Performance Evaluation of Speech Recognition under Noisy Conditions. In: ISCA ITRW ASR 2000, Paris, pp. 181–188 (2000)Google Scholar