基于改进K中心点的语音分组识别算法

Grouping voice recognition algorithm based on improved K center point

针对语音分组识别中,分组数量多时,识别率下降,分组数量少时,识别时间增加这一问题,提出一种改进K中心点的语音分组识别算法。该算法先将同一语音进行分段均值降维,然后将不同语音经过K中心点聚类分组。识别时先判断所在语音组,再进行模式匹配。实验结果表明,相比于传统K均值聚类和传统K中心点算法,该算法能减少识别时间,提高识别率。

In speech group recognition,when the number of groups is large,the recognition rate decreases,while when the number of groups is small,the recognition time increases.Therefore a speech grouping recognition algorithm with improved K center point was proposed.The algorithm first will be the same speech segment,K⁃means,dimension reduction,then different speeches are grouped through K⁃center point clustering.In recognition,the speech group is identified first,and then pattern matching is carried out.Experimental results show that compared with traditional K⁃means clustering and traditional K center point algorithm,this algorithm can reduce the recognition time and improve the recognition rate.