A robust begin-end point detector for highly noisy conditions

Most recognition methods, which have shown to be highly efficient under noise-free conditions fail dramatically with S/N ratios around or below 10 dB. One of the consequences of these high noise levels is that most Begin-End Point Detectors fail to separate properly the speech segments of the noise ones. Therefore, the speech recognition mechanisms will not have a clear boundary to start the processing of the signal, and as a consequence, speech segments will be lost, and noisy segments will be used in recognition. The overall reliability of the Speech Recognition System will dramatically experience the consequences of this impairment in its results. What is being proposed in this paper is to use the side information provided by an Adaptive Noise Canceller for the dynamic detection of word boundaries.