Tutorial on image reconstruction based on weighted sum (WS) filter approach: from single image to multi-frame image

Abstract

For large magnification factors, the prior classical smoothness leads to overly smooth results with very little high-frequency content. The classical image restorations are failing to reconstruct the desired image. Consequently, the Recognition-Based Restoration is desired for these purposes and one of the most effective techniques is of a weighted sum (WS) filter. This paper reviews the research framework of weighted sum (WS) filter approach for image reconstruction. This research framework first starts with the Hard-Partition-based Weighted Sum (HP-WS) filter proposed in 1999 and then consequently reviews the Subspace HP-WS (S-HPWS) using PCA (Principal Component Analysis) Filter proposed in 2005, The Soft-Partition-based Weighted Sum (SP-WS) proposed in 2006, and the fast Adaptive Wiener Filter proposed in 2007. The paper reviews each filter technique in terms of its computational concepts, demonstrates parameter optimization from the point of view of the mathematical analysis, and discusses advantages and disadvantages.