专利名称：Method of speech recognition using time-dependent interpolation and hiddendynamic value classes
发明人：Li Deng,Jian-lai Zhou,Frank Torsten Bernd
Seide,Asela J. R. Gunawardana,HagaiAttias,Alejandro Acero,Xuedong Huang
申请号：US11294858申请日：20051206公开号：US07206741B2公开日：20070417
专利附图：
摘要：A speech signal is decoded by determining a production-related value for acurrent state based on an optimal production-related value at the end of a precedingstate, the optimal production-related value being selected from a set of continuousvalues. The production-related value is used to determine a likelihood of a phone beingrepresented by a set of observation vectors that are aligned with a path between thepreceding state and the current state. The likelihood of the phone is combined with ascore from the preceding state to determine a score for the current state, the scorefrom the preceding state being associated with a discrete class of production-relatedvalues wherein the class matches the class of the optimal production-related value.
申请人：Li Deng,Jian-lai Zhou,Frank Torsten Bernd Seide,Asela J. R. Gunawardana,HagaiAttias,Alejandro Acero,Xuedong Huang
地址：Redmond WA US,Beijing CN,Beijing CN,Seattle WA US,Seattle WA US,BellevueWA US,Bellevue WA US
国籍：US,CN,CN,US,US,US,US
代理机构：Westman, Champlin & Kelly, P.A.
代理人：Theodore M. Magee
更多信息请下载全文后查看
因篇幅问题不能全部显示，请点此查看更多更全内容