Broad phoneme class recognition in noisy environments using the GEMS

Broad phoneme class recognition has the advantage of offering additional acoustic-phonetic knowledge to the speech processing applications. In several papers, exploiting such information is shown to be advantageous for HMM-based speech enhancement systems. The problem with those systems is the dramatic decrease in recognition accuracy in noisy environments. In this work, we extract the energy feature from an auxiliary sensor and directly fuse it with the features extracted from the speech signal. Experiment results with noisy speech show significant increase in performance.