基于深度学习的声纹识别系统优化研究  

Research on Optimization of Voiceprint Recognition System Based on Deep Learning

作　　者：贾雅晴 JIA Yaqing(Hefei University of Economics,Hefei 230012,China)

机构地区：[1]合肥经济学院,安徽合肥230012

出　　处：《电声技术》2024年第3期33-35,共3页Audio Engineering

摘　　要：聚焦于基于深度学习的声纹识别系统优化方法,重点探讨了堆叠循环神经网络(Rerrent Neural Network,RNN)模型在声纹识别中的应用。首先介绍了基于深度学习的声纹识别系统的基本架构,其次引入堆叠RNN模型作为优化方法,最后在MATLAB平台上利用VoxCeleb数据集进行实验验证。实验结果表明,相比于标准RNN模型,堆叠RNN模型在准确率、精确率、召回率及F1值等评价指标上均取得了显著提高,验证了该方法在声纹识别任务中的有效性和优越性。This paper focuses on the optimization method of voiceprint recognition system based on deep learning,and focuses on the application of Rerrent Neural Network(RNN)model in voiceprint recognition.Firstly,the basic architecture of voiceprint recognition system based on deep learning is introduced.Secondly,the stacked RNN model is introduced as an optimization method.Finally,experiments are carried out on the platform of MATLAB using VoxCeleb data set.The experimental results show that,compared with the standard RNN model,the stacking RNN model has significantly improved the accuracy,precision,recall and F1 value,which verifies the effectiveness and superiority of this method in voiceprint recognition.

关 键 词：深度学习 声纹识别 循环神经网络(RNN) 堆叠优化 

分 类 号：TN912.34[电子电信—通信与信息系统]