Data-driven temporal filters based on multi-eigenvectors for robust features in speech recognition.

Ni-Chun Wang, Jeih-Weih Hung, Lin-Shan Lee

Browse the full ICASSP paper archive.