A Dynamic Multi-exposure Image Fusion Method based on Image Registration

Abstract

In this paper, a novel multi-exposure image fusion method is proposed. Different from the previous MEF method, this method can fuse both static multi-exposure images and dynamic multi-exposure images. Most of the existing dynamic multi-exposure image fusion methods focus on the removal of ghosting artifacts, while ignoring the color and details of the fusion image. Rich images. The motion detector is used to identify the moving part of the image, obtain the image sequence of different exposure levels consistent with the motion of the reference image, convert the multi-exposure image of the dynamic scene into the multi-exposure image of the static scene, and then obtain the final fusion image through the fusion model. The proposed method has good performance in both subjective and objective evaluation.