首页>
外国专利>
METHOD TO SIMPLIFY AN SVM-BASED SPEECH AND MUSIC CLASSIFIER FOR SELECTABLE MODE VOCODER CODEC BASED ON CONTRIBUTIONS OF SUPPORT VECTOR
METHOD TO SIMPLIFY AN SVM-BASED SPEECH AND MUSIC CLASSIFIER FOR SELECTABLE MODE VOCODER CODEC BASED ON CONTRIBUTIONS OF SUPPORT VECTOR
展开▼
机译:基于支持向量贡献的基于模式支持向量机的语音和音乐分类器选择方法的简化方法
展开▼
页面导航
摘要
著录项
相似文献
摘要
The present invention relates to a kind of simplified methods to support voice and music classifier based on machine, simplifying method supports voice and the music classifier classifier according to the present invention that is easy to implement based on machine to enter embedded system by reducing calculation amount, while selectable modes audio coder (SMV) of category support vector machines (SVM) classifier of retention property based on background model. In accordance with the invention it is possible to have SVM-based classifiers by the SMV codecs of the SVM-based background models of simplifying support vector classifier reducing to operate behind which, it is easy that especially realization classifier, which enters embedded system,. ;The 2014 of copyright KIPO submissions;[Reference numerals] (AA) calculates the vector of average contribution; (BB) relativce contribution vectors are calculated; (CC) removal support vector machines has smaller contribution compared with threshold value
展开▼