Taylor series and adaptive directional selection for real time demosaicking

摘要

In this paper, we introduce an efficient demosaicking method based on Taylor series and adaptive direction selection using adaptive weight. Taylor series is used to approximate the to-be-interpolated value using values of neighbor pixels in four directions. Instead of estimating the edge orientations via limited candidate directions as conventional demosaicking methods, we propose an adaptive direction and weight selection based demosaicking method to achieve high efficiency. The weight is used to evaluate the similarity among the neighbor pixels. Experimental results indicate that the proposed method outperforms existing approaches in both objective and subjective performance.