Multi-Exposure Image Fusion Through Feature Decomposition

Abstract

Multi-exposure image fusion is an effective method for fusing differently exposed low dynamic range (LDR) images to a high dynamic range (HDR) image. The previous methods suffer from poor detail and color restoration performance and visual artifact, such as halo. In this paper, to overcome these problems, we propose a novel network architecture for multi-exposure image fusion (MEF) based on feature decomposition and RGB channel fusion. A feature of LDR image is decomposed to the common and residual components at a feature level. Then, fusion is performed on the respective common and residual domain. It is found through diverse experiments that the proposed network could improve the MEF performance in aspects of color restoration and visual artifact.