摘要
语音识别技术已在通信及控制等领域得到广泛应用,针对孤立词语音识别矢量量化中LBG算法对初始码书选择敏感,容易陷入局部最优、泛化能力不强的缺点,将免疫粒子群优化算法(IPSO)和LBG算法结合进行聚类分析,从而得到基于IPSO-LBG的码书设计方法,并将其用于基于离散隐马尔可夫模型(DHMM)的孤立词语音识别系统中。通过实验,与传统LBG算法的DHMM孤立词语音识别系统的识别结果相比,证明了改进的系统有较好的识别率和适应性。
Abstract:Speech recognition technology is widely applied in many fields, such as communication, automatic control etc. In isolated word speech recognition based on DHMM, we must quantify the feature vector using LBG algorithm. In this paper, LBG algorithm trained by immune particle swarm optimization algorithm is proposed in order to overcome disadvantage of LBG algorithm which heavily relies on the initial code book, and has local optimum and weak adaptive capacity. Experimental results demonstrate that the improved algorithm has good performance, which is better than the speech recogrfition b^sed on DIqMM by T,13C, ~lgorlthrn in the recoonnirinn rote anti mhll^rneg~
出处
《数字技术与应用》
2013年第1期111-113,共3页
Digital Technology & Application
关键词
孤立词识别免疫粒子群优化LBG算法DHMM
isolated word speech recognition LBG algorithm immune particle swarm optimization algorithm DHMM