Research Area:  Machine Learning
Single image super-resolution (SISR) is one of the contemporary research areas in the field of image restoration that involves solving an ill-posed inverse equation. A rich profusion of techniques has been proposed in the past four decades. However, the expansion of deep learning (DL) in recent years has improved image reconstruction drastically. In this paper, a compendium of DL applications in the SISR field has been provided, mainly focusing on inspection of the latest advancements and categorization. For completeness of the study, different aspects of DL based SISR are included in brief with a synoptic study on the available image datasets. In conclusion, issues existing in DL-based SISR and viable solutions for them are proffered.
Keywords:  
Deep Learning
Single-Image Super-Resolution
Machine Learning
Author(s) Name:  Garima Pandey & Umesh Ghanekar
Journal name:  Pattern Recognition and Image Analysis
Conferrence name:  
Publisher name:  Springer
DOI:  10.1134/S1054661822010059
Volume Information:  volume 32, pages11–32 (2022)
Paper Link:   https://link.springer.com/article/10.1134/S1054661822010059