×

Segmentation approach for speech recognition systems

  • US 6,535,851 B1
  • Filed: 03/24/2000
  • Issued: 03/18/2003
  • Est. Priority Date: 03/24/2000
  • Status: Expired due to Term
First Claim
Patent Images

1. A method for automatically determining a set of phonetic units from a body of utterance data, the method comprising the computer-implemented steps of:

  • receiving the body of utterance data;

    determining a first set of candidate phonetic units from the body of utterance data;

    determining a set of no-cross regions from the body of utterance data wherein the no-cross regions correspond to a time span of utterance data having a high probability of containing a boundary between phonetic units;

    filtering the first set of candidate phonetic units to generate a subset of candidate phonetic units therefrom wherein the filtering analyzes the candidate phonetic units to determine if the candidate spans a no-cross region for the utterance data such that the subset omits candidate phonetic units which spanned a no-cross region.

View all claims
  • 5 Assignments
Timeline View
Assignment View
    ×
    ×