遗传优化RBF神经网络背景下的声纹识别研究  

Research on Voiceprint Recognition Based on Genetic Optimized RBF Neural Network

作　　者：张鹏[1] ZHANG Peng(Hubei Police Officer Academy,Wuhan Hubei 430034,China)

机构地区：[1]湖北警官学院,湖北武汉430034

出　　处：《信息与电脑》2022年第1期159-161,共3页Information & Computer

基　　金：法庭科学湖北省重点实验室(湖北警官学院)资助课题“一种基于深度学习的声纹特征采样方法研究”(项目编号:KFKT2021005)。

摘　　要：本文以遗传优化RBF神经网络为基础分析声纹识别算法,通过增加学习网络权重能力和全局寻优能力,提升声纹识别准确度。对语音进行预处理并提出语音特征参数,以遗传RBF神经网络为基础,从初始化种群、选择适应度函数、选择操作、交叉和变异操作4个方面完成模型构建。通过仿真试验分析,得到遗传优化RBF神经网络背景下声纹识别的技术优势。This paper analyzes the voiceprint recognition algorithm based on the genetic optimization RBF neural network, and improves the accuracy of voiceprint recognition by increasing the learning network weight ability and global optimization ability. The speech is preprocessed and the speech characteristic parameters are proposed. Based on the genetic RBF neural network, the model is constructed from four aspects: initializing the population, selecting the fitness function, selecting operation, crossover and mutation operation. Through simulation test analysis, the technical advantages of voiceprint recognition under the background of genetically optimized RBF neural network are obtained.

关 键 词：遗传算法 RBF神经网络 声纹识别 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]