RelUNet: Relative Channel Fusion U-Net for Multichannel Speech Enhancement
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917796211851264 |
|---|---|
| author | Aldarmaki, Ibrahim Solorio, Thamar Raj, Bhiksha Aldarmaki, Hanan |
| author_facet | Aldarmaki, Ibrahim Solorio, Thamar Raj, Bhiksha Aldarmaki, Hanan |
| contents | Neural multi-channel speech enhancement models, in particular those based on the U-Net architecture, demonstrate promising performance and generalization potential. These models typically encode input channels independently, and integrate the channels during later stages of the network. In this paper, we propose a novel modification of these models by incorporating relative information from the outset, where each channel is processed in conjunction with a reference channel through stacking. This input strategy exploits comparative differences to adaptively fuse information between channels, thereby capturing crucial spatial information and enhancing the overall performance. The experiments conducted on the CHiME-3 dataset demonstrate improvements in speech enhancement metrics across various architectures. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_05019 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | RelUNet: Relative Channel Fusion U-Net for Multichannel Speech Enhancement Aldarmaki, Ibrahim Solorio, Thamar Raj, Bhiksha Aldarmaki, Hanan Sound Machine Learning Audio and Speech Processing Neural multi-channel speech enhancement models, in particular those based on the U-Net architecture, demonstrate promising performance and generalization potential. These models typically encode input channels independently, and integrate the channels during later stages of the network. In this paper, we propose a novel modification of these models by incorporating relative information from the outset, where each channel is processed in conjunction with a reference channel through stacking. This input strategy exploits comparative differences to adaptively fuse information between channels, thereby capturing crucial spatial information and enhancing the overall performance. The experiments conducted on the CHiME-3 dataset demonstrate improvements in speech enhancement metrics across various architectures. |
| title | RelUNet: Relative Channel Fusion U-Net for Multichannel Speech Enhancement |
| topic | Sound Machine Learning Audio and Speech Processing |
| url | https://arxiv.org/abs/2410.05019 |