An Enhanced Lossless Color Filter Array Image Compression Based On Predictive Adaptive Arithmetic Coding

Abstract

Most consumer digital cameras use a single image light sensor which provides color information using color filter array(CFA).This  provided a mosaic images, in which each pixel position contains only one color component in case of Bayer CFA Pattern. This paper produced a CFA hierarchical prediction scheme based on context adaptive coding. In CFA hierarchical scheme, the green pixels were subdivided into two sets .  was encoded by a gray scale conventional method and  was  predicted based on . The red pixels were predicted using both the sets of green pixels and blue pixels were predicted using red and green. The predictors were designed based on direction of the edges in the neighborhood. Using the prediction information, the magnitude of prediction error was also determined and context adaptive arithmetic coding was applied to reduce bits. The simulated results on CFA images showed that the proposed method gives less bits per pixel than the recently developed CFA compression algorithms.