| Due to the huge gap between the high dynamic range of natural scenes and the limited range of consumer-grade cameras,a single-shot image can hardly record all the information of a scene.Multi-exposure image fusion(MEF)has been an effective way to solve this problem by integrating multiple shots with different exposures,which is in nature an enhancement problem.This thesis mainly studies the MEF task,and proposes two MEF methods.The specific research contents and main contributions of this thesis are summarized as follows:(1)Focusing on the analysis and optimization of the shortcomings of existing work,this thesis proposes an MEF method based on information richness measurement.From the perspective of measuring the information richness of the fused image and restricting the correlation between the fused image and the source images,this method designs a set of self-supervised loss functions,and proposes an efficient fusion network structure,so as to realize high-quality multi exposure image fusion.Experimental results show that the proposed method can achieve good fusion results for different kinds of scenes.(2)This thesis reanalyzes the characteristics of multi-exposure fusion task,and draws a conclusion that during fusion,informativeness and the visual realism should be concerned simultaneously.To achieve the goal,this thesis presents a deep perceptual enhancement network for MEF.Specifically,the proposed method contains two modules,one of which responds to gather content details from inputs while the other takes care of color mapping/correction for final results.Both extensive experimental results and ablation studies are conducted to show the efficacy of proposed method,and demonstrate its superiority over other state-of-the-art alternatives both quantitatively and qualitatively.This thesis also verifies the flexibility of the proposed strategy on improving the exposure quality of single images.Moreover,the proposed method can fuse 720 p images in more than 60 pairs per second on an Nvidia 2080 Ti GPU,making it attractive for practical use. |
