Automatic speech recognition system and method

G - Physics – 10 – L

Patent

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

G10L 15/08 (2006.01)

Patent

CA 2592861

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).

L'invention concerne un procédé de reconnaissance vocale automatique qui permet d'identifier les mots à partir d'un signal vocal d'entrée, y compris de prévoir au moins une reconnaissance par hypothèse fondée sur le signal vocal d'entrée, la reconnaissance d'hypothèses étant un mot d'hypothèses individuel ou une série de mots d'hypothèses individuels, et de calculer une mesure de confiance (CDC) de la reconnaissance d'hypothèses, sur la base du signal vocal d'entrée. Le calcul d'une mesure de confiance (CDC) comporte le calcul de contributions différentielles (CI,t) à la mesure de confiance (CDC), chacune comme différence entre un score acoustique contraint (SCVi,t) et un score acoustique non contraint (SCLt), la pondération de chaque contribution différentielle (CI,t) par application d'une fonction de distribution cumulative (F.xi.I) de la contribution différentielle (CI,t) de façon à rendre la distribution des mesures de confiance (CDC) homogènes en termes de fonctionnalité de rejet, en fonction des variations de langage, de vocabulaire et de grammaire, et calcul de la mesure de confiance (CDC) par établissement de la moyenne des contributions différentielles pondérées (CI,t).

LandOfFree

Say what you really think

Search LandOfFree.com for Canadian inventors and patents. Rate them and share your experience with other people.

Rating

Automatic speech recognition system and method does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Automatic speech recognition system and method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Automatic speech recognition system and method will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFCA-PAI-O-1736480

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.