S3R-Net: A Single-Stage Approach to Self-Supervised Shadow Removal
This work addresses shadow removal for computer vision applications, offering a non-cyclic, unidirectional self-supervised approach that is incremental compared to existing supervised and self-supervised methods.
The paper tackles the problem of shadow removal in images by proposing S3R-Net, a self-supervised network that uses a two-branch WGAN to unify style and adapt from unaligned shadow-free references, achieving comparable numerical scores to recent self-supervised models with superior qualitative performance and low computational cost.
In this paper we present S3R-Net, the Self-Supervised Shadow Removal Network. The two-branch WGAN model achieves self-supervision relying on the unify-and-adaptphenomenon - it unifies the style of the output data and infers its characteristics from a database of unaligned shadow-free reference images. This approach stands in contrast to the large body of supervised frameworks. S3R-Net also differentiates itself from the few existing self-supervised models operating in a cycle-consistent manner, as it is a non-cyclic, unidirectional solution. The proposed framework achieves comparable numerical scores to recent selfsupervised shadow removal models while exhibiting superior qualitative performance and keeping the computational cost low.