A Self-Learning Approach to Edge-Adaptive Demosaicking Algorithm for CFA Images

Most of the current digital cameras will capture only one color at each of the pixel using a single sensor overlaid with a color filter array (CFA).Inorder to recover a full color image from a mosaic pattern, specific methods will be used to restore the two missing color values at each of the pixels in the Bayer pattern.The restoration process to find out the missing color samples is known as color demosaicking. Color-difference based edge-adaptive filtering will reduce the error in the red and blue channel by exploiting the high frequency information of the green channel in the proposed enhanced edge-adaptive demosaicking algorithm. By using the self-learning approach via support vector regression, it extracts the image dependent information in constructing the learning model. It does not require any additional training data externally. Hence the discuss method will shows that the proposed system will provide state off the art in both subjective and objective image quality. Keywords—Digital camera,Color demosaicking, color filter array,adaptive filtering,color artifacts,selflearning,support vector regression.