CCSR-Net: Unfolding Coupled Convolutional Sparse Representation for Multi-focus Image Fusion

Multi-focus image fusion aims to generate an all-in-focus image from multiple partially focused images of the same scene captured with different focal settings. In this paper, we present a coupled convolutional sparse representation (CCSR) model for multi-focus image fusion. Instead of being solved by an iterative thresholding algorithm, the proposed CCSR model is unfolded into a learnable neural network (termed as CCSR-Net) using the deep unfolding technique, taking the advantages of both traditional methods and deep-learning (DL)-based ones. Based on the CCSR-Net, a new multi-focus image fusion method with good interpretability is further proposed. Experimental results on two popular datasets show that the proposed method can obtain the state-of-the-art performance in terms of both visual quality and objective assessment.