Research on Voiceprint Recognition Based on Improved Densenet

In the study of voiceprint recognition, the model parameters are large, the robustness is poor, and the features of voiceprint are not fully utilized. In this paper, an improved method based on Densenet framework for voice print recognition is proposed. Using DenseNet as backbone network, SE attention module is used to improve Dense Block in DenseNet network. The experimental results show that the improved Densenet (ERR: 0.87^%) is superior to the traditional voice print recognition network.