Edge-based adaptive support region computing algorithm for local tone mapping

Conventional dodging-and-burning in Reinhard's photographic algorithm often suffer from halo artifacts or a loss of details, which is due to the failure of distinguishing the regions with few strong edges from the ones with lots of weak details during the local mapping. In this paper, we propose an edge-based adaptive support region computing algorithm for local tone mapping. The proposed algorithm operates maximum filters with gradually extending size for the edge of the original image to compute support regions, which manages to make a distinction between weak details and strong edge effectively, and could get more appropriate support regions. Experiments show that the proposed algorithm can not only preserve more details but also run much faster.