Image Fusion Algorithm Using Pyramidal Empirical Mode Decomposition

Multi-scale decomposition (MSD) approaches are very useful in image processing and play an important role in image fusion algorithm. This paper proposes a novel image fusion algorithm using pyramidal empirical mode decomposition (PEMD). The principle of PEMD consists of performing a pyramid transform on intrinsic mode functions (IMF) and the residual image of empirical mode decomposition (EMD). The input images are decomposed into a sequence of detail pyramidal images at different levels of resolution and an approximation image by PEMD. Fusion algorithm is performed on the input images decompositions to produce the composite PEMD representation and then the inverse PEMD transform is applied to obtain the fused image. The experimental results show that the fusion algorithm by using PEMD gives more encouraging and effective performance than EMD and traditional pyramid fusion algorithms.