mlsanityData to 7 Oct 2026

MUSE: Flexible Voiceprint Receptive Fields and Multi-Path Fusion Enhanced Taylor Transformer for U-Net-based Speech Enhancement

arXiv 2406.04589 ↗First posted 7 Jun 2024cs.SD cs.IR cs.ITcited by 4

Not on the map. Why?

Abstract (arXiv)

Achieving a balance between lightweight design and high performance remains a challenging task for speech enhancement. In this paper, we introduce Multi-path Enhanced Taylor (MET) Transformer based U-net for Speech Enhancement (MUSE), a lightweight speech enhancement network built upon the Unet architecture. Our approach incorporates a novel Multi-path Enhanced Taylor (MET) Transformer block, which integrates Deformable Embedding (DE) to enable flexible receptive fields for voiceprints. The MET Transformer is uniquely designed to fuse Channel and Spatial Attention (CSA) branches, facilitating channel information exchange and addressing spatial attention deficits within the Taylor-Transformer framework. Through extensive experiments conducted on the VoiceBank+DEMAND dataset, we demonstrate that MUSE achieves competitive performance while significantly reducing both training and deployment costs, boasting a mere 0.51M parameters.

Key facts. Cited by 4 papers in the map (data through 7 Oct 2026). Builds on 23 of its 32 references that are in the map. First posted on arXiv 7 Jun 2024.

Builds on

23 of 32 references.

References · 23

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  4. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,827
  5. 2022Simple Baselines for Image Restorationcited by 701
  6. 2019ECA-Net: Efficient Channel Attention for Deep Convolutional Neural Networkscited by 572
  7. 2017SEGAN: Speech Enhancement Generative Adversarial Networkcited by 165
  8. 2021MetricGAN+: An Improved Version of MetricGAN for Speech Enhancementcited by 114
  9. 2023MB-TaylorFormer: Multi-branch Efficient Transformer Expanded by Taylor Formula for Image Dehazingcited by 81
  10. 2019An Empirical Study of Spatial Attention Mechanisms in Deep Networkscited by 58
  11. 2019Demucs: Deep Extractor for Music Sources with extra unlabeled data remixedcited by 53
  12. 2023MP-SENet: A Speech Enhancement Model with Parallel Denoising of Magnitude and Phase Spectracited by 51
11 more
  1. 2019PHASEN: A Phase-and-Harmonics-Aware Speech Enhancement Networkcited by 49
  2. 2022CMGAN: Conformer-based Metric GAN for Speech Enhancementcited by 45
  3. 2021TSTNN: Two-stage Transformer based Neural Network for Speech Enhancement in the Time Domaincited by 36
  4. 2018Differentiable Consistency Constraints for Improved Deep Speech Enhancementcited by 35
  5. 2021DPT-FSNet: Dual-path Transformer Based Full-band and Sub-band Fusion Network for Speech Enhancementcited by 33
  6. 2020Speech Enhancement using Self-Adaptation and Multi-Head Self-Attentioncited by 28
  7. 2022MANNER: Multi-view Attention Network for Noise Erasurecited by 28
  8. 2021Dual-branch Attention-In-Attention Transformer for single-channel speech enhancementcited by 18
  9. 2022A Two-Stage U-Net for High-Fidelity Denoising of Historical Recordingscited by 17
  10. 2022Multi-View Attention Transfer for Efficient Speech Enhancementcited by 6
  11. 2023Efficient Encoder-Decoder and Dual-Path Conformer for Comprehensive Feature Learning in Speech Enhancementcited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 1; 2025: 2; 2026: 1.

4 papers.

  1. 2024Mamba-SEUNet: Mamba UNet for Monaural Speech Enhancementcited by 10
  2. 2025Exploring Resolution-Wise Shared Attention in Hybrid Mamba-U-Nets for Improved Cross-Corpus Speech Enhancementcited by 0
  3. 2025IMSE: Efficient U-Net-based Speech Enhancement using Inception Depthwise Convolution and Amplitude-Aware Linear Attentioncited by 0
  4. 2026BASENet: Band-Adapted Speech Enhancement Network with Cross-Band Attentioncited by 0