Image Super Resolution Using ESRGAN: A Preliminary Experimental Study

The ongoing progression of technology has led to substantial advancements in image and video quality. Elevated image and video quality inherently enhance visual experiences. However, not all visual content maintains a high standard. Diverse factors contribute to the degradation of image and video quality. Nevertheless, super-resolution techniques offer a solution to restore compromised quality in images and videos. Super-resolution involves enhancing low-resolution images into high-resolution versions, often achieved through deep learning-based methodologies. This research focuses on utilizing the Enhanced Super-Resolution Generative Adversarial Network (ESRGAN) to elevate image quality, providing the potential for generating images of superior quality from low-resolution sources.

I. Introduction

Along with technological advances, the quality of images and videos continues to undergo significant improvement. Higher image and video quality undoubtedly result in more captivating visual experiences. However, not all images or videos maintain high-quality attributes. Various factors can contribute to image and video quality deterioration, such as image or video compression processes. Nevertheless, the compromised quality of images or videos can be rectified by employing super-resolution techniques.