Shadow Detection And Removal From Images

Abstract

Shadow removal from images and videos is an essential task in computer vision that concentrates on detecting the shadow generated by the obstructed light source and obtaining realistic shadow free results. The features of shadows are the same as those of objects. As a result, it has the potential to be misclassified as a part of the object, resulting in degrading the performance of many computer vision tasks. In recent years, several deep learning based frameworks have been presented to solve this issue. This work presents a method based on Generative Adversarial Networks (GANs) for shadow removal by supervised learning. Specifically, we train two generators and two discriminators to learn the mapping between shadow and shadow free domains. We employ generative adversarial constraints with cycle consistency and content constraints to learn the mapping efficiently. We also propose an adaptive exposure correction module to handle the over exposure problem in the shadow area of the result. We additionally present a method for improving the quality of benchmark datasets and eventually achieving better shadow removal results. We also show ablation studies to analyze the importance of the ground truth data with the adaptive exposure correction module in the proposed framework and explore the impact of using different learning strategies in the presented method. We validate the proposed approach on the two large scale upervised benchmark datasets and show quantitative and visual improvements in the state of the art results.