Shadow removal of single texture region using local histogram matching

摘要

Shadows have long been a problem to computer vision algorithms. Removing shadow can significantly improve the performance of several vision task such as object detection and image segmentation. Various methods of shadow removal in images had been developed. In this paper, we discussed the characteristic of shadowed region with single texture statistically, based on an illumination model, and then developed a simple and fast way of removal shadow using local histogram matching in different illumination reduction level.