Voice activity detection using frame-wise model re-estimation method based on Gaussian pruning with weight normalization.

Masakiyo Fujimoto, Shinji Watanabe, Tomohiro Nakatani

Browse the full Interspeech paper archive.