Related Experiment Video
Updated: Jan 16, 2026
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
1.0K
Comprehensive Review of Deep Learning Approaches for Single-Image Super-Resolution
Zirun Liu1,2, Shijie Jiang3, Shuhan Feng3
1Longmen Laboratory, Luoyang 471000, China.
Sensors (Basel, Switzerland)
|September 27, 2025
Summary
Deep learning-based single-image super-resolution (SISR) enhances image resolution by overcoming system limitations. This review provides a framework for understanding SISR
Area of Science:
- Computer Vision
- Image Processing
- Deep Learning
Background:
- Single-image super-resolution (SISR) addresses physical limitations in imaging systems to improve resolution.
- Deep learning methods have become central to advancing SISR techniques.
Purpose of the Study:
- To systematically introduce deep learning-based SISR methods.
- To propose a method-oriented classification framework for SISR.
- To explore theoretical basis, technological evolution, and domain-specific applications of SISR.
Main Methods:
- A method-oriented classification framework is proposed.
- Key technical components like datasets, upsampling strategies, objective functions, and quality assessment are analyzed.
- Classic SISR model reconstruction results are compared.
Main Results:
- A systematic knowledge framework for SISR is presented.
- In-depth analysis of benchmark datasets, multi-scale upsampling, and objective function optimization.
- Comparative evaluation of established SISR models.
Conclusions:
- The review provides a comprehensive overview of deep learning-based SISR.
- Identifies limitations and proposes future research directions for SISR.
- Offers significant reference value for the advancement of SISR technology.
