LRetUNet: A U-Net-based retentive network for single-channel speech enhancement

Yuxuan Zhang, Zipeng Zhang, Weiwei Guo, Wei Chen, Zhaohai Liu, Houguang Liu

Abstract: Highlights•This work presents the procedure for implementing a speech enhancement method.•A U-Net structure with a retentive mechanism for STFT-based speech enhancement.•We designed the hidden filter module to the skip connections in the U-Net architecture.•The method outperforms baselines with fewer parameters and Multiplies-Accumulate Ops..