Multi-exposure Dynamic Image Fusion Based on PatchMatch and Illumination Estimation

In this study, we present a novel image fusion algorithm for multi-exposure dynamic images based on PatchMatch and illumination estimation. To eliminate the ghosting artifacts which often occur in the fusion results of existing exposure fusion methods when there are moving objects in the scenes, the fusion process of our proposed algorithm is as follows. First, we take advantage of the PatchMatch method to align the selected reference image with the other input images and then we fuse these images together based on illumination estimation to obtain the final fusion image. Experimental results demonstrate that our proposed method performs better than the existing fusion methods both in visual effect and objective indicators.