A Novel Multi-exposure Image Fusion Approach Based on Parameter Dynamic Selection
This paper propose a parameter dynamic selection approach for multi-exposure image fusion (MEF) that based on image cartoon-texture and structural patch decomposition. The image texture component is obtained by using texture-cartoon decomposition from the input image. The dynamic parameter is achieved by calculating the image texture entropy. The image patch is divided into three conceptually independent components by using structural patch decomposition. Respectively processing and fusing these three components, a fusion patch and aggregate fused patches are reconstruct into a fused image. This novel MEF method achieves dynamic parameter selection by utilizing texture-cartoon decomposition to obtain fusion images with more details.
KeywordsMulti-exposure image fusion High dynamic range imaging Parameter dynamic selection Structural patch decomposition