A Novel Image Enhancement Algorithm Based on Information Fusion

This paper presented a novel image enhancement algorithm using information fusion (fusionEnhance) to preserve image edges in the image filtering. The proposed algorithm firstly defined fuzzy-based region entropy inspired by the concept of local entropy and region direction gradient respectively as information fusion sources, and then it implemented the triangle module fusion operation for image edge detection. Besides the fusion-decision edge detection, the proposed algorithm operated a modified hybrid filter method to preserve image edges for image enhancement in image filtering. Experiment results illustrate that fusionEnhance can improve the enhancement performance in terms of PSNR and two fusion metrics-image entropy and edge similarity (indicating edge preserving performance), compared with AMWM.