Feature generation based on maximum normalized acoustic likelihood for improved speech recognition.

Xiang Li, Richard M. Stern

Browse the full ICASSP paper archive.