An Iterative Super-Resolution Reconstruction of Image Sequences using a Bayesian Approach with BTV prior and Affine Block-Based Registration

Abstract

The traditional SR image registrations are based on translation motion model therefore super-resolution applications can apply only on the sequences that have simple translation motion. In this paper, we present a novel image registration, the fast affine block-based registration, for performing super-resolution using multiple images. We propose super-resolution reconstruction that uses a high accuracy registration algorithm, the fast affine block-based registration [15], and is based on a maximum a posteriori estimation technique by minimizing a cost function. The L1 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 errors due to possibly inaccurate motion estimation. Bilateral regularization is used as prior knowledge for removing outliers, resulting in sharp edges and forcing interpolation along the edges and not across them. The experimental results show that the proposed reconstruction can apply on real sequence such as Suzie.