Efficient speech enhancement based on left-right HMM with state sequence detection using LRT

Since the conventional HMM (Hidden Markov Model)-based speech enhancement methods try to improve speech quality by considering all states for the state transition, hence introduce huge computational loads inappropriate to real-time implementation. In the Left-Right HMM (LR-HMM), only the current and the next states are considered for a possible state transition so to reduce the computation complexity. We propose a new speech enhancement algorithm based on LR-HMM with state sequence detection using LRT (Likelihood Ratio Test). Experimental results show that the proposed method improves the speed up with little degradation of speech quality compared to the conventional method.