A noise-robust ASR back-end technique based on weighted viterbi recognition.

Xiaodong Cui, Alexis Bernard, Abeer Alwan

Browse the full Interspeech paper archive.