5th European Conference on Speech Communication and Technology

Rhodes, Greece
September 22-25, 1997

Voice Activity Detection Using Source Separation Techniques

Tomohiko Taniguchi, Shoji Kajita, Kazuya Takeda, Fumitada Itakura

Dept. of Information Electronics, Graduate School of Engineering, Nagoya University, Nagoya, Japan

Blind signal separation method based on minimiz- ing mutual information is applied to deal with multi- speaker problem in speech recognition. Recognition experiments performed under di erent acoustic environments, in a soundproof room and a reverberant room, clarify that 1) the method can improve recognition accuracy by about 20% where SNR condition is 0 dB, 2) the method is more effective when many speakers' speech exist than the simple overlapped situation, and that 3) the method does not work well under reverberant conditions.

Full Paper

Bibliographic reference.  Taniguchi, Tomohiko / Kajita, Shoji / Takeda, Kazuya / Itakura, Fumitada (1997): "Voice activity detection using source separation techniques", In EUROSPEECH-1997, 1103-1106.