Cross-attention among spectrum, waveform and SSL representations with bidirectional knowledge distillation for speech enhancement
Authors:
- Hang ChenHCHang Chen
- Chenxi WangCWChenxi Wang
- Qing WangQWQing Wang
- Jun DuJDJun Du
- Sabato Marco SiniscalchiSSSabato Marco Siniscalchi
- Genshun WanGWGenshun Wan
- Jia PanJPJia Pan
- Huijun DingHDHuijun Ding
- 24 April 2025
- journal article
- Published by Elsevier in Information Fusion
- Vol. 122
Abstract
No abstract availableFunding Information
- National Natural Science Foundation of China
- University of Science and Technology of China (62171427)
This publication has 58 references indexed in Scilit:
- A novel target decoupling framework based on waveform-spectrum fusion network for monaural speech enhancementDigital Signal Processing, 2023
- Real-time neural speech enhancement based on temporal refinement network and channel-wise gating methodsDigital Signal Processing, 2022
- Joint waveform and magnitude processing for monaural speech enhancementApplied Acoustics, 2022
- Glance and gaze: A collaborative learning framework for single-channel speech enhancementApplied Acoustics, 2022
- Correlating subword articulation with lip shapes for embedding aware audio-visual speech enhancementNeural Networks, 2021
- Fullsubnet: A Full-Band and Sub-Band Fusion Model for Real-Time Single-Channel Speech Enhancement2021
- ICASSP 2021 Deep Noise Suppression Challenge: Decoupling Magnitude and Phase Optimization with a Two-Stage Deep Network2021
- SDR – Half-baked or Well Done?2019
- Stochastic Optimization Methods2015
- The distance between two random vectors with given dispersion matricesLinear Algebra and its Applications, 1982