| 摘要 |
<p>Disclosed herein is an automatic speech recognition method for identifying words from an input speech signal, including providing at least one hypothesis recognition based on the input speech signal, the hypothesis recognition being an individual hypothesis word or a sequence of individual hypothesis words, and computing a confidence measure (C DC ) for the hypothesis recognition, based on the input speech signal, wherein computing a confidence measure (C DC )includes computing differential contributions (C i,t) to the confidence measure (C DC ), each as a difference between a constrained acoustic score (SCV i t ) and an unconstrained acoustic score (SCL t ),weighting each differential contribution (Ci,t) by applying thereto a cumulative distribution function (F.xi. i )of the differential contribution (C i , t ),so as to make the distributions of the confidence measures (C DC )homogeneous in terms of rejection capability, as the language, vocabulary and grammar vary, and computing the confidence measure (C DC )by averaging the weighted differential contributions (Ci,t).</p> |
