DBHN-Net: Dual-Branch Hybrid Network for Speech Enhancement

Arxiv CS.SD (Sound/Audio ML)

The paper proposes DBHN-Net, a dual-branch hybrid neural network for low-complexity monaural speech enhancement. It aims to reduce computational cost while maintaining high performance for practical deployment.

speech-enhancement

Arxiv CS.SD (Sound/Audio ML)·Cunhang FanEnrui LiuJing ZhouJian KangJie LiAndong LiJian ZhouZhao LvXuelong Li

6 days ago