Multi-focus Image Fusion Based on NSST and Focus Region Detection

Abstract

In this work, a new multi-focus image fusion method based on non-subsampled Shearlet Transform (NSST) and focus area detection is proposed. This method effectively utilizes the information of multiple sources to obtain an all-in-focus image. The source images are firstly transformed by NSST to obtain the low frequency coefficients and high frequency coefficients. Then the boundary map and weight map are obtained by focus region detection from the high frequency coefficients. Next, we use the NSST-based image fusion method and the spatial-domain method based on the focus region detection to get the fused images, respectively. Finally, the all-in-focus image can be constructed by the fused images with boundary map. Experiment results show that the proposed method outperforms the other fusion methods. It can make the focus and defocus regions fused more naturally without boundary effects.