基于元音长度调整及分组的中文孤立词识别  

Recognition of Single Chinese Character Based on the Modulation of Length and Grouping of Vowel

作　　者：朱力[1] 曾毓敏[1] 汪育春[1] 葛倩蓉[1] 

机构地区：[1]南京师范大学物理科学与技术学院,南京210046

出　　处：《信息化研究》2011年第4期14-17,共4页INFORMATIZATION RESEARCH

摘　　要：大量统计结果表明,语速的快慢,在一定程度上,会影响语音识别率。为了解决这个问题,在原有动态时间规整基础上,论文采用了元音长度调整算法,先找出元音部分的一个完整周期,再通过插入或者删除相似波形,以达到调整语速的效果。论文提出了元音分组算法,分别对元音按照类型、相似性以及长度进行分组。给不同的分组使用不同的阈值进行计算,以提高识别率。Many statistical results show that, to some extent, speech rate will affect recognition probability. To solve this problem, based on the original DTW algorithm, the article uses an algorithm, adjustment of the length of vowel. First of all, we will find a whole cycle of vowel part. Then, we insert or remove the similar waveform to adjust the speed. Moreover the article introduced the vowel grouping algorithm, such as grouping based on the type of vowel, grouping based on similarity, grouping based on length. Different grouping with different threshold value are computed to improve the reorganization rate.

关 键 词：中文孤立词 动态时间规整 元音长度调整 语速调整 元音分组 

分 类 号：TN912.3[电子电信—通信与信息系统]