Hard C-means clustering for voice activity detection

Abstract

An effective voice activity detection (VAD) algorithm is proposed for improving speech recognition performance in noisy environments. The proposed speech/pause discrimination method is based on a hard-decision clustering approach built on a set of subband log-energies and noise prototypes that define a cluster. Detecting the presence of speech (a new cluster) is achieved using a basic sequential algorithm scheme (BSAS) according to a given “distance” (in this case, geometrical distance) and a suitable threshold. The accuracy of the Cluster VAD (ClVAD) algorithm lies in the use of a decision function defined over a multiple-observation (MO) window of averaged subband log-energies and a suitable noise subspace model defined in terms of prototypes. In addition, the reduced computational cost of the clustering approach makes it adequate for real-time applications, i.e. speech recognition. An exhaustive analysis is conducted on the Spanish SpeechDat-Car databases in order to assess the performance of the proposed method and to compare it to existing standard VAD methods. The results show improvements in detection accuracy over standard VADs such as ITU-T G.729, ETSI GSM AMR and ETSI AFE and a representative set of recently reported VAD algorithms for noise robust speech processing.