An effective integer demosaicing based on directional filtering

Nowadays many digital still cameras and mobile phones use single image sensor to capture the colors of a scene. In this way, only one of the three primary colors is obtained at each pixel location and the other two are missing. The goal of demosaicing or color interpolation is to estimate the missing colors so that a full-color image can be reconstructed. Although many demosaicing algorithms have been proposed, few of them are suitable for the implementation of integer (or fixed-point) processors and give good performance. To solve this problem, we present an effective integer demosaicing technique based on directional filtering. An appealing property is that all the divisions in the technique can be implemented by using only shifts. Moreover, a new directional refining step is proposed to improve the reconstructed image quality. Experimental results show that the proposed technique can give better performance than several integer techniques.