Multi-wavelet guided deep mean-shift prior for image restoration

作者：

Highlights：

• The model inherits structural representation in wavelet and learning ability in CNN.

• Multi-channel features generated by multi-wavelets is rearranged as the network input.

• It involves the recurrent structure-preserving and multi-view aggregation properties.

摘要

•The model inherits structural representation in wavelet and learning ability in CNN.•Multi-channel features generated by multi-wavelets is rearranged as the network input.•It involves the recurrent structure-preserving and multi-view aggregation properties.

论文关键词：Image restoration,Deep mean-shift prior,Proximal gradient,Wavelet transform,Recurrent structure-preserving,Multi-view complementary aggregation

论文评审过程：Received 18 November 2020, Revised 29 June 2021, Accepted 23 August 2021, Available online 5 September 2021, Version of Record 15 September 2021.

论文官网地址：https://doi.org/10.1016/j.image.2021.116449