一种基于语音组成单位的说话人识别算法.PDF

第21卷摇 第6 期摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 模式识别与人工智能摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 Vol.21摇 No.6 摇 2008年12月摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 PR AI摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 Dec摇 摇 2008 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 一一种基于语音组成单位的说话人识别算法 黄长存摇 摇 汪增福 (中国科学技术大学 自动化系合肥摇 230027) 摘摇 要摇 以线性预测系数为特征通过高斯混合模型的迭代算法对训练样本的初始 均值聚类结果进行优化,得到k 语音组成单位的表示.以语音组成单位的模式匹配为基础,提出一种文本无关说话人确认的方法———均值法,以及 一种文本无关说话人辨认方法.实验结果表明,即使在短时语音下本文方法都能取得较好效果. 关键词摇 语音组成单位,文本独立说话人确认,文本独立说话人辨认 + 中图法分类号摇 TP242.6 3 A Speaker Recognition Algorithm Based on Speech Component Unit HUANG Chang鄄Cun,WANGZeng鄄Fu (Department of Automation,University of Science and Technology of China,Hefei230027) ABSTRACT The linear prediction coefficients (LPC) are selected as features to construct the corresponding feature space. Firstly,the samples for training of each individual in system are clustered in the feature space toform initial clustersby usingk鄄meanclustering algorithm. Then,the initial clusteringresultsare optimized based on Gaussian mixture model (GMM) iterative algorithm to obtain the speech component unit representation of the individual. On the basis of the obtained speech component units, a text鄄 independent speaker verification method, called averaging method, and a text鄄independent speaker identification method arepresented. Experimental resultsshow that theproposed algorithm canproduce a satisfying result even in short utterances. Key Words摇 Speech Component Unit, Text鄄Independent Speaker Verification, Text鄄Independent Speaker Identification 1摇 引摇 言 题是,所抽取的语音特征受说话内容的制约,不具有