A Novel empirical mode decomposition based system for medical image enhancement

Abstract

In this paper, we introduce a system for enhancing medical images. The proposed system utilizes Ensemble Empirical Mode Decomposition (EEMD) to decompose the signal into distinct frequency components called intrinsic mode functions (IMFs). These components will be enhanced individually and then recombined to construct the enhanced image. The novelty of the proposed approach is in the method of enhancement and combination of the IMFs. The experimental results demonstrate the performance of the proposed algorithm in visualizing many details that are hidden in the original image. Compared with some existing methods, such as Histogram Equalization, LSD ACE, cascaded unsharp masking and tile-based local enhancement, the new method shows to be more effective in enhancing the images that consist of varying illumination in several regions.