Fast monotonic blind deconvolution algorithm for constrained TV based image restoration

A new fast monotonic blind deconvolution algorithmic method is investigated based on the constrained variational minimization framework under the periodic boundary conditions. The contributions of our methodology are that the blur operator identification and image restoration can be simultaneously optimized even under high noise level as compared to previous methods. Specifically, the monotone fast iterative shrinkage/thresholding algorithm (MFISTA) combined with the fast gradient projection (FGP) algorithm, is extended to deal with our new proposed algorithm and guarantee the monotonic convergence rate. In addition, the deblurring subproblem is enhanced by incorporating a bisection technique to effectively identify a near optimal value for the regularization parameter of the TV-Frobenius objective function quickly and accurately. Initial experimental results for gray satellite and color wireless capsule endoscopy (WCE) images demonstrate the considerable performance of the proposed algorithm.