Multi channel echo cancellation and speech enhancement for embedded systems method
Speech enhancement has become a front-end processing technology in contemporary society.In order to meet the growing demand for communication and explore better and less complex speech enhancement systems,a multi-channel echo cancellation speech enhancement method for embedded systems has been designed,which combines the UNet network structure and channel attention mechanism,Two echo cancellation algorithm models have been proposed:the MP-UNet model and the lightweight version of the MP-UNets model.In both models,the amplitude and phase of speech can be processed simultaneously,and speech that eliminates echoes and suppresses noise can be output.The experimental verification of the performance superiority of the proposed method has significant implications for speech enhancement systems in resource constrained embedded platform development environments.
万方数据
维普
