基于HMM和WNN的心音信号身份识别研究*

将隐马尔可夫模型（HMM）与小波神经网络（WNN）相结合，提出了一种基于心音信号的身份识别方法。该方法首先利用HMM对心音信号进行时序建模，并计算出待识别心音信号的输出概率评分；再将此识别概率评分作为小波神经网络的输入，通过小波神经网络将HMM的识别概率值进行非线性映射，获取分类识别信息；最后根据混合模型的识别算法得出识别结果。实验采集80名志愿者的160段心音信号对所提出的方法进行验证，并与GMM模型的识别结果进行了对比，结果表明，所选方法能够有效提高系统的识别性能，达到了比较理想的识别效果。

Study of human identity recognition based on HMM and WNN

Used HMM combined with WNN, this paper proposed a human identity recognition method based on heart sound signal. First, employed HMM to train the time sequence of heart sounds and to compute the Viteribi output score. Then used the score as the input of WNN, made nonlinear mapping by WNN to acquire the classification information. The result of recognition was made by these two kinds of recognition information. The experiment collected 160 heart sounds from 80 people to test the proposed algorithm, the proposed system achieved higher recognition rate than Gaussian mixture model method. The results show that the hybrid model can effectively improve the recognition performance of the system and achieve a satisfactory effect.