A Tukey’s biweight bayesian approach for a robust iterative SRR of image sequences

The SRR (super-resolution reconstruction) performance degrades severely when the data model assumptions do not faithfully describe the measure data. Therefore, the robust estimator applicable to nonfaithful data models is desired in SRR algorithms. This paper proposes an alternate SRR approach to solve 2 major causes of performance degradation in conventional SRR approaches. The two causes are (1) an invalid assumption of data model and (2) a registration error in registration stage. Our proposed SRR is based on the stochastic regularization technique of Bayesian MAP estimation by minimizing a cost function. The Tukey's biweight norm is used for measuring the difference between the projected estimate of the high-resolution image and each low resolution image, removing outliers in the data and Tikhonov regularization is used to remove artifacts from the final answer and improve the rate of convergence. The experimental results show that the proposed SRR can apply on real sequence such as Suzie and Foreman sequence and confirm the effectiveness of our method and demonstrate its superiority to other SRR methods based on L1 and L2 norm for a several noise models such as AWGN, Poisson and speckle noise.