On Restoration and Super-Resolution for Continuous Image Sequence - Adaptive Filtering Approach

Among the existing algorithms for the restoration of continuous image sequences and estimation of super-resolution images, there is no unified approach which progressively generates a restored super-resolution output image sequence from a given low resolution sequence. Moreover, typical video applications may require coping with linear time and space variant blur and arbitrary motion, which none of the existing algorithms support. This paper presents a new method based on adaptive filtering theory for restoration and super-resolution image estimation. The new approach generalizes the super-resolution algorithms, merged with the image sequences restoration techniques. The proposed methodology suggests least squares (LS) filters which adapt in time, based on the LMS, RLS and the Kalman Filtering algorithms. The adaptation enables the treatment of linear space and time variant blurring and arbitrary motion, both of them assumed known. The proposed algorithms are modified by adding a regul..