Feature transformation based on discriminant analysis preserving local structure for speech recognition.

Makoto Sakai, Norihide Kitaoka, Kazuya Takeda

Browse the full ICASSP paper archive.