噪声环境中基于GMM汉语说话人识别

Analysis of GMM To Chinese Speaker Identification In Noisy Environments

在噪声环境中如何提高说话人识别精度是说话人识别研究中的一个非常重要的课题。为了让说话人识别系统在安静的环境和噪声环境中都能取得满意的工作性能,研究一个将语音增强器与说话人识别器级连起来的系统。该系统中,采用Weiner滤波增强算法提高说话人识别系统前端预处理的抗噪声能力,提高输入信号的信噪比。实验测试结果表明,该系统具有很好的抗噪声性能。

This paper presents our recent work on a noisy speaker identification system. It is important thesis how to enhance the fight ratio in speaker identification.We studied a speaker identification system that can work well both in noisy environment and quiet environment.In this system,we adopt Weiner filter algorithm to improve the Signal-Noise Ratio of the input signal for speaker identification. Experimental results show that the system is satisfactory in noisy environments.