Guided Image Filter Using Multi-Exposure Image Fusion: A Patch-Wise Approach

Abstract

A lot of research has been contributed in the area of optimization of design. MOST applications in computer vision and computer graphics involve image filtering to suppress and/or extract content in images. We propose a patch-wise approach for multi-exposure image fusion (MEF) using guided image filter. A key step in our approach is to decompose each colour image patch into three conceptually independent components: signal strength, signal structure and mean intensity. Upon processing the three components separately based on patch strength and exposedness measures, we uniquely reconstruct a colour image patch and place it back into the fused image. Unlike most pixel-wise MEF methods in the literature, the proposed algorithm does not require significant pre/post processing steps to improve visual quality or to reduce spatial artefacts. Moreover, the novel patch decomposition allows us to handle RGB colour channels jointly and thus produces fused images with more vivid colour appearances using guided image filter. Extensive experiments demonstrate the superiority of the proposed algorithm both qualitatively and quantitatively with guided filter. Hybrid thresholding is applied on remaining sub bands (LH, HL and HH).The proposed work will be implemented using MATLAB R2015a.