UnShadowNet: Illumination Critic Guided Contrastive Learning For Shadow Removal

TL;DR

UnShadowNet is a novel weakly supervised framework for shadow removal that leverages contrastive learning and illumination guidance, outperforming state-of-the-art methods on multiple datasets.

Contribution

It introduces a new weakly supervised shadow removal method using contrastive learning and illumination guidance, with an extension to fully-supervised training.

Findings

Outperforms existing methods on ISTD, adjusted ISTD, SRD datasets

Effective in both weakly and fully supervised setups

Achieves superior shadow removal quality

Abstract

Shadows are frequently encountered natural phenomena that significantly hinder the performance of computer vision perception systems in practical settings, e.g., autonomous driving. A solution to this would be to eliminate shadow regions from the images before the processing of the perception system. Yet, training such a solution requires pairs of aligned shadowed and non-shadowed images which are difficult to obtain. We introduce a novel weakly supervised shadow removal framework UnShadowNet trained using contrastive learning. It is composed of a DeShadower network responsible for the removal of the extracted shadow under the guidance of an Illumination network which is trained adversarially by the illumination critic and a Refinement network to further remove artefacts. We show that UnShadowNet can be easily extended to a fully-supervised set-up to exploit the ground-truth when available. UnShadowNet outperforms existing state-of-the-art approaches on three publicly available shadow datasets (ISTD, adjusted ISTD, SRD) in both the weakly and fully supervised setups.