ENHANCED IMAGE RESTORATION THROUGH FUSION: BLUR ROI APPROACH

Abstract

Image fusion is the technique of joining two or more images for best conditional awareness. Fusing process takes the important feature from one image and combines with best feature of other images. This experiment present a pixel level image fusion of multi-focus images based on the two-stage fuzzy fusion. In this approach, we first applied two different fusion approaches on the input image, and then applied fuzzy fusion approach on the output of the previous stage. Performance of this approach, output of fusion image is degraded where focused area in source image and good to blur area of source image. Blur are two types: defocus blur which is caused by imaging systems and motion blur caused by the relative motion between camera and objects. The proposed approach is applicable on the defocused blur images, images are taken in the focused and blur region segmentation by Local Binary Pattern (LBP) based and then find intersection blur region of source image, then finally applied two stage fuzzy fusion approach on blur region. And output of fusion result is combined to focused area of source images. The experiment result of proposed approach is enhances image quality compare to other available fusion approach.