HMM retraining based on state duration alignment for noisy speech recognition.

Wei-Wen Hung, Hsiao-Chuan Wang

Browse the full Interspeech paper archive.