Ninth International Conference on Spoken Language Processing

Pittsburgh, PA, USA
September 17-21, 2006

Noisy Speech Recognition Based on Selection of Multiple Noise Suppression Methods Using Noise GMMs

Norihide Kitaoka, Souta Hamaguchi, Seiichi Nakagawa

Toyohashi University of Technology, Japan

To achieve high recognition performance for a wide variety of noise and for a wide range of signal-to-noise ratio, this paper presents integration methods of four noise reduction algorithms: spectral subtraction with smoothing of time direction, temporal domain SVD-based speech enhancement, GMM-based speech estimation and KLT-based comb-filtering. In this paper, we proposed two types of ?combination methods of noise suppression algorithms: selection of front-end processor and combination of results from multiple recognition processes. Recognition results on the AURORA-2J task showed the effectiveness of our proposed methods.

Full Paper

Bibliographic reference.  Kitaoka, Norihide / Hamaguchi, Souta / Nakagawa, Seiichi (2006): "Noisy speech recognition based on selection of multiple noise suppression methods using noise GMMs", In INTERSPEECH-2006, paper 1207-Thu2CaP.12.