Data-driven temporal filters for robust features in speech recognition obtained via Minimum Classification Error (MCE).

Jeih-Weih Hung, Lin-Shan Lee

Browse the full ICASSP paper archive.