Benchmarking Adversarial Robustness of Image Shadow Removal with Shadow-adaptive Attacks

arXiv:2403.10076v1 Announce Type: new
Abstract: Shadow removal is a task aimed at erasing regional shadows present in images and reinstating visually pleasing natural scenes with consistent illumination. While recent deep learning techniques have demonstrated impressive performance in image shadow removal, their robustness against adversarial attacks remains largely unexplored. Furthermore, many existing attack frameworks typically allocate a uniform budget for perturbations across the entire input image, which may not be suitable for attacking shadow images. This is primarily due to the …

abstract adversarial adversarial attacks arxiv attacks benchmarking consistent cs.cv deep learning deep learning techniques image images natural performance regional robustness shadow type