摘要
基于特征选择的语音特征获取用于说话人识别是目前较为有效的方式。但是,最优语音特征随着具体应用环境的变化而不同。因此,提出了基于四类型语音特征封装式遗传特征选择算法(FSF-Wr GAF),该算法提取了四种类型的语音特征参数,通过链式智能体遗传算法和GMM-UBM进行封装式动态特征选择,获取高精度的识别准确率。采用了多种指标完成该算法的性能测试。实验结果表明,该算法具体实现过程简便,改进效果明显,较同类算法在多项指标(识别率,EER,DET曲线)上都有显著提高。
Speech feature extraction based on feature selection is a very effective method for speaker recognition. However, the optimal speech features have also changed. Therefore, this paper proposes a kind of four kinds of speech feature wrapper selection framework algorithm(FSF-WrGAF). The algorithm extracts four kinds of speech features, and conducts dynamic wrapper feature selection by Chainlike Agent Genetic Algorithm(CAGA)and Gaussian Mixture Model-Universal Back-ground Model(GMM-UBM), thereby obtaining high recognition accuracy. Several algorithms are compared in the experiment part. Experimental results show that the FSF-WrGAF algorithm can obtain apparent improvement in terms of accuracy, equal error rate and detection cost compared with some other algorithms.
出处
《计算机工程与应用》
CSCD
北大核心
2016年第14期150-155,219,共7页
Computer Engineering and Applications
基金
国家自然科学基金(No.91438104)
中央高校基本科研业务费专项资金(No.CDJZR10160003
No.CDJZR13160008
No.CDJZR155507)
中国博士后科学基金(No.2013M532153)
重庆市博士后科研项目特别资助