Voice recognition system using implicit speaker adaption

Voice recognition system using implicit speaker adaption

  • CN 101,221,759 A
  • Filed: 03/22/2002
  • Published: 07/16/2008
  • Est. Priority Date: 03/28/2001
  • Status: Active Application
First Claim
Patent Images

1. audio recognition method, described method comprises the steps:

  • The first input sound bite is carried out pattern match with at least the first template, to produce the pronunciation kind that at least one input pattern coupling is estimated and determined to be identified;

    Follow the corresponding evaluation relevant with at least the second template to make comparisons described at least one input pattern coupling evaluation, wherein said at least the second template is from the language person certain acoustic model relevant with the pronunciation kind of being discerned;

    WithAccording to described comparative result, determine whether to upgrade described at least the second template.

View all claims
    ×
    ×

    Thank you for your feedback

    ×
    ×