A particular directional multilevel transform based method for single-image rain removal

作者：

Highlights：

• An uni-directional multilevel system is constructed to describe the inherent sparse property of the rainy image, rain and clean image.

• A convex model for single image rain removal problem is proposed, which includes two multilevel transform spare regularizers, a sparse regularizer of rain streaks and the fidelity term.

• The split Bregman algorithm is used to solve the model, its computation includes three soft-thresholding processes, two fast Fourier transforms and its inverse, the computation complexity is .

摘要

•An uni-directional multilevel system is constructed to describe the inherent sparse property of the rainy image, rain and clean image.•A convex model for single image rain removal problem is proposed, which includes two multilevel transform spare regularizers, a sparse regularizer of rain streaks and the fidelity term.•The split Bregman algorithm is used to solve the model, its computation includes three soft-thresholding processes, two fast Fourier transforms and its inverse, the computation complexity is .

论文关键词：Rain removal,Directional multilevel transform,Split Bregman algorithm

论文评审过程：Received 16 July 2019, Revised 22 April 2020, Accepted 1 May 2020, Available online 14 May 2020, Version of Record 18 May 2020.

论文官网地址：https://doi.org/10.1016/j.knosys.2020.106000