Reference-free low-light image enhancement by associating hierarchical wavelet representations

作者：

Highlights：

• The idea of hierarchical wavelet transform with a highway connection is proposed.

• The highway connection is a residual-like connection for different frequency bands.

• We ingeniously introduces frequency bands into the reference-free loss function.

• Experimental results demonstrate the superiority of the proposed method.

摘要

•The idea of hierarchical wavelet transform with a highway connection is proposed.•The highway connection is a residual-like connection for different frequency bands.•We ingeniously introduces frequency bands into the reference-free loss function.•Experimental results demonstrate the superiority of the proposed method.

论文关键词：Image enhancement,Low-light,Deep learning

论文评审过程：Received 16 May 2022, Revised 4 September 2022, Accepted 25 September 2022, Available online 30 September 2022, Version of Record 12 October 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.118920