FullSubNet

Platform Python version Pytorch Version GitHub repo size

This Git repository for the official PyTorch implementation of “A Full-Band and Sub-Band Fusion Model for Real-Time Single-Channel Speech Enhancement”, submitted to ICASSP 2021.

?
See the demo page in this link.

workflow

fullsubnet_result

You can use all of these things:

  • Available models
    • FullSubNet
    • Delayed Sub-Band LSTM
    • Fullband LSTM Baseline
  • Available Datasets
    • Deep Noise Suppression Challenge – INTERSPEECH 2020
    • Demand + CSTR VCTK Corpus

Documentation

Citation

If you use this code for your research, please consider citing:

@misc{hao2020fullsubnet,
      title={FullSubNet: A Full-Band and Sub-Band Fusion Model for Real-Time Single-Channel Speech Enhancement}, 
      author={Xiang Hao and Xiangdong Su and Radu Horaud and Xiaofei Li},
      year={2020},
      eprint={2010.15508},
      archivePrefix={arXiv},
      primaryClass={eess.AS}
}

License

License: MIT

GitHub

https://github.com/haoxiangsnr/FullSubNet