Method for updating voiceprint feature model and terminal

Method for updating voiceprint feature model and terminal

  • CN 102,760,434 A
  • Filed: 07/09/2012
  • Published: 10/31/2012
  • Est. Priority Date: 07/09/2012
  • Status: Active Application
First Claim
Patent Images

1. vocal print characteristic model update method is characterized in that said method comprises:

  • Obtain the raw audio streams that comprises at least one speaker;

    Speaker according to preset is cut apart and clustering algorithm, obtains described in the said raw audio streams other audio stream of each speaker'"'"'s branch among at least one speaker;

    Other audio stream of branch of each speaker among said at least one speaker is mated with original vocal print characteristic model respectively, obtain the successful audio stream of coupling;

    The audio stream that said coupling is successful appends the audio stream training sample as the said original vocal print characteristic model of generation, and said original vocal print characteristic model is upgraded.

View all claims
    ×
    ×

    Thank you for your feedback

    ×
    ×