摘要 |
<p>An input utterance containing a command word to be recognized is processed (110) and features which adequately represent the utterance are determined. Prestored features of a set of reference samples of command words (160) are compared (170) to the features of the input utterance. Recognition of command words in noisy environments is improved by determining the distance between the features of the input utterance and the features of the reference samples and modifying the distance (120) in response to background noise. The reference sample having the minimum distance is selected as the recognized command word.</p> |