Multi-focus Image Fusion based on Edge-preserving Filters

To break the limitation of camera imaging and acquire abundant information with multi-focus images, we present a novel multi-focus image fusion method based on edge-preserving filters. In this paper, the focusing level is measured by two cost functions and the focus map is constructed based on the winner- take-all manner. Besides, we demonstrate that the guided image filter (GIF) and the fast global smoother (FGS) have different advantages in image structure transferring and image smoothing, which can be utilized to construct the precise fusion weight maps. Combing with the weight maps acquired by GIF and FGS, the accurate all-in-focus image is obtained using a secondary fusion strategy. Experimental results show that the proposed method is competitive or even outperforms many state-of-the-art methods, which includes the recent CNN-based fusion method, while the proposed method is less time-consuming.