A Hybrid Method for Multi-Focus Image Fusion Based on Fast Discrete Curvelet Transform

This paper presents a fast discrete Curvelet transform (FDCT)-based technique for multi-focus image fusion to address two problems: texture selection in FDCT domain and block effect in spatial-based fusion. First, we present a frequency-based model by performing FDCT on the input images. Considering the human visual system characteristics, a union of pulse coupled neural network and sum-modified-Laplacian algorithms are proposed to extract the detailed information of frequencies. Then, we construct a hybrid spatial-based model. Unlike other spatial-based methods, we combine the image difference and the detailed information extracted from input images to detect the focused region. Finally, to evaluate the robustness of proposed method, we design a completed evaluation process considering the misregistration, noise error, and conditional focus situations. Experimental results indicate that the proposed method improves the fusion performance and has less computational complexity compared with various exiting frequency-based and spatial-based fusion methods.