Image Super-Resolution: A Complete Guide

Image/video super-resolution are research thrust areas in recent times. Their applications include HDTV, image coding, image resizing, image manipulation, remote sensing, face recognition, astronomy, and surveillance. The objective is to increase image/ video resolution through upsampling, deblurring, denoising, deep learning etc. The development of various image/ video super-resolution theories has been studied in this book focusing on Deep convolutional networks-based super-resolution (DeepCNSR). More than 30 state-of-the-art super-resolution Convolutional Neural Networks (CNNs) over three classical and three recently introduced challenging datasets to benchmark single image super-resolution have been exhaustively analyzed with its merits and demerits. A taxonomy with nine categories for DeepCNSR networks has been introduced including linear, residual, multi-branch, recursive, progressive, attention-based, and adversarial designs. Network complexity, memory footprint, model input and output, learning details, the type of network losses, and important architectural differences (e.g., depth, skip-connections, filters) of each model have been studied comparatively.