Spatio-temporal regularization featuring novel temporal priors and multiple reference motion estimation

Abstract

Handling input content ranging from low resolution and highly compressed internet streams up to high definition video is a challenging task for modern flat panel TVs. Considering that the consumer is accustomed to HD content, superior processing is mandatory especially for low bit-rate video with very low quality. In this paper, we describe a novel spatio-temporal and image content adaptive regularization algorithm, which can deal with this entire content well. This algorithm can significantly improve the video quality, remove spatial and temporal artifacts and even re-establish missing details. The spatio-temporal regularization combines the well-known Total Variation image model with a novel temporal constraint, being robust towards erroneous motion vectors and allowing application of block based motion estimation techniques.

A New Adaptive Video Super-Resolution Algorithm With Improved Robustness to Innovations

“…Most works in single-frame or video SRR seek robustness by considering cost functions including non-quadratic (e.g. L 1 ) error norms [4], [7], [25] or signal dependent regularizations [9], [36], which result in non-linear algorithms. Although these techniques achieve good reconstruction results, their increased computational cost makes real-time operation unfeasible even for the fastest algorithms.…”