A new speech/non-speech classification method using minimal Walsh basis functions

In this paper, a method to distinguish between the closely located speech and non-speech segments of noisy speech signals is presented. It is based on modification of the original signal with the help of minimal binary Walsh basis functions and an analysis/synthesis scheme. Then, the presence of speech/non-speech segments is determined from the modified signals with a simple decision scheme. Besides its simplicity, the simulation results are promising and significantly better than G.729B with signal to noise ratio (SNR) as low as 0 dB in different noisy environments.

[1]  Ming Zhang,et al.  A robust speech detection algorithm in a microphone array teleconferencing system , 2001, 2001 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.01CH37221).

[2]  John Riley,et al.  Exploiting Walsh-based attributes to stereo vision , 1996, IEEE Trans. Signal Process..

[3]  Wonyong Sung,et al.  A statistical model-based voice activity detection , 1999, IEEE Signal Processing Letters.

[4]  Moncef Gabbouj,et al.  Wavelet-based corner detection technique using optimal scale , 2002, Pattern Recognit. Lett..

[5]  Joon-Hyuk Chang,et al.  Voice activity detection based on complex Laplacian model , 2003 .

[6]  A. Kondoz,et al.  Analysis and improvement of a statistical model-based voice activity detector , 2001, IEEE Signal Processing Letters.

[7]  K. Chon,et al.  A Robust Time-Varying Identification Algorithm Using Basis Functions , 2003, Annals of Biomedical Engineering.