A Novel Medical Image Fusion Method Based on MCA and MOPSO

Abstract

Medical image fusion on different modalities has provided new diagnostic information for clinical medicine. To further enhance its effect, we present a new medical image fusion approach based on morphological component analysis (MCA) and multi-objective particle swarm optimization (MOPSO). In this method, MCA separates the source images into texture images and cartoon images by employing the morphological diversity of source images. Moreover, MOPSO is adopted to optimize the parameters in the fusion rules. For cartoon images, region energy and MOPSO are utilized to obtain more detailed information from the separated cartoon images. For texture images, the larger texture image coefficients are selected by sum-modified-Laplacian (SML) scheme. Finally, the fused image can be reconstructed by weighting the fused texture image and cartoon image with MOPSO. The experimental results indicate that the fused image by the proposed method outperforms many other medical image fusion methods in both visual effect and objective evaluation.