Abstract
Image super-resolution (SR) has emerged as a pivotal technique in computer vision, enhancing low-resolution images to higher resolutions for diverse applications such as medical imaging, satellite imagery, and surveillance. Traditionally, signal processing methods have been employed for SR, leveraging interpolation and reconstruction algorithms. However, recent advancements in machine learning (ML), especially deep learning, have revolutionized SR by learning complex mappings from low-resolution to high-resolution images. This paper presents a comprehensive overview of the integration of machine learning and classical signal processing approaches in image super-resolution. We discuss the strengths and limitations of both domains, explore hybrid methodologies, and highlight their applications. Furthermore, experimental comparisons demonstrate how combining ML models with signal processing techniques improves reconstruction quality, computational efficiency, and robustness. The study concludes with future directions for optimized hybrid SR frameworks.

This work is licensed under a Creative Commons Attribution 4.0 International License.
Copyright (c) 2021 Dr. Elena V. Petrova (Author)