Image restoration using alpha map-based fusion for multiple color aperture computational camera

We present a novel image restoration approach by fusing multiple images acquired by a multiple color aperture computational camera for overcoming the limit of conventional image restoration algorithms, such as high computational load and post-processing artifacts. The proposed restoration framework utilizes color matching, alignment, and fusion of multiple images acquired by using multiple color apertures. The alpha map-based image fusion step successfully restores high-frequency components of meaningful edge regions. The proposed scheme can achieve automatic focusing and varying depth of filed at low cost for low-end consumer cameras.