A VTS-based feature compensation approach to noisy speech recognition using mixture models of distortion.

Jun Du, Qiang Huo

Browse the full ICASSP paper archive.