Application of improved U-Net network with attention mechanism in end-to-end speech enhancement  

作　　者：WU Ruiqin CHEN Xueqin YU Jie WANG Lirong ZHAO Heming 

机构地区：[1]School of Electronic and Information Engineering,Soochow University,Suzhou 215006

出　　处：《Chinese Journal of Acoustics》2022年第4期390-403,共14页声学学报（英文版）

基　　金：supported by the National Natural Science Foundation of China(61340004)。

摘　　要：An improved U-Net(Attention Dilated Convolution U-Net,ADC-U-Net)network model for end-to-end speech enhancement is designed based on the U-Net network.Compared with the baseline U-Net network,the dilated convolution is added to reduce the loss of information caused by sampling.Besides,the attention mechanism structure is introduced,which combines more contextual information of noisy speech to extract deeper and richer feature information.The proposed model avoids the extraction of features with distinct step,so it does not need three steps of traditional methods,including feature extraction,feature denoising and waveform reconstruction.The proposed network model obtains complex structural features to represent speech through multi-level and multi-scale learning.The quality and intelligibility of enhanced speech are evaluated by several subjective and objective indexes.Experimental data show that the proposed algorithm performs well in noise suppression and adaptability,and has advantage over baseline U-Net network and other models in speech quality and intelligibility.

关 键 词：NETWORK SPEECH MECHANISM 

分 类 号：TN912.35[电子电信—通信与信息系统]