An Efficient CFA Image Reconstruction Algorithm using Sparse Computation

Compressive sensing can reduce the number of measurements required for display of an image thereby, reducing storage space and balance the traffic through data communication networks. In this paper a frequency domain algorithm for efficient reconstruction of CFA Bayer filter images has been proposed. Frequency transformations will be achieved by using standard wavelet decomposition algorithms. Sparse values of CFA output image, in transformed domain will be computed using prediction algorithms (l1 magic-number gradient algorithm). Effectiveness of the algorithm will be computed using CPSNR, PSNR, MSE and IEF. This algorithm is expected to improve the visual quality of the CFA images.