ATF-MUSE: A Lightweight U-Net-Based Model With Adaptive Time-Frequency Transformer for Speech Enhancement

Read full text Summarise with VIRAT
Authors: Yongsheng Yu, Zhuoran Cai, Hao Jiang
Cited by: 0
References: 31

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Speech enhancement techniques based on deep learning have demonstrated substantial improvements over traditional signal processing methods, particularly under complex non-stationary noise conditions. In this paper, we propose a novel lightweight speech enhancement model, ATF-MUSE, which is designed to more effectively model the complex interactions between temporal- and frequency-domain features, while dynamically adjusting attention weights to enhance adaptability to varying noise environments. Specifically, we introduce an ATF-Transformer module that alternately models dependencies along the temporal and frequency dimensions, thereby capturing both the global spectral patterns and long-term temporal relationships of speech signals. Moreover, we propose an Adaptive Multi-Head Attention (AMHA) mechanism that dynamically generates feature-dependent attention weights, significantly improving the model’s robustness against non-stationary noise. We evaluate the proposed ATF-MUSE model on the widely used VoiceBank-DEMAND dataset, and experimental results demonstrate that our method outperforms several state-of-the-art approaches across multiple standard evaluation metrics, including PESQ, MOS, and STOI.

Fields of Study & Subjects

Speech and Audio Processing Speech Recognition and Synthesis Image and Signal Denoising Methods

Publication details

Journal / source
IEEE Access
Volume
14
Pages
46294–46303
Publication date
1 January 2026
Language
EN