Image enhancement and post-processing for low-resolution compressed video

Abstract

This paper presents an adaptive two-stage algorithm for obtaining high-resolution video from a compressed low-resolution video. In the first stage, an adaptive projection onto convert algorithm for removing coding artifacts from a low-resolution compressed video is described. A convex set is defined using quantization step size and the degree of discontinuity between neighboring pixels. In the second stage, a regularization approach is applied for obtaining a high-resolution video from a low-resolution video, from which the coding artifacts have been removed; here, the point spread function for obtaining high-resolution frames is estimated by backward projection. Experimental results validate the proposed algorithm's performance. © 2014 International Information Institute.