Adaptive Design of Color Filter Arrays by using the Frequency Domain Approach

Abstract

In digital color imaging, the image is obtained by a single sensor which is extend over by a color filter array (CFA), in every pixel only single color component has to measure by using single sensor. Demosaicking is the procedure of rebuild the complete color image from a raw image. So the CFA can be effected by irreversible visual artifacts, the quality of the demosaicked images is decisive for the CFA and demosaicking algorithm. Providentially, to control these issues the CFA design in frequency domain allocates a theoretical approach. Basically, in the frequency domain all conventional methods there is an involvement of human beings. In this paper, we are introducing adaptive design of CFA in the frequency domain. In this method the mosaicked images mainly depends on the frequency structure representation. To present the frequency structure candidates we are using a multi-objective optimization, by using this overlapping between the mosaicked frequency image components with the CFA is minimized. In the constrained optimization problem each candidate parameters are optimized. The technique used to solve the optimization problem is alternating direction method (ADM). Arbitrary frequency structure is the application of parameter optimization method, including the conjugate replicas of chrominance components. The advantages of the proposed method for the benchmark images are shown in experimental results. In the future work, bilateral filter is used to improve the performance as well as quality of an image with increased SNR value.