Thermal Image Super-Resolution Challenge Results - PBVS 2022

This paper presents results from the third Thermal Image Super-Resolution (TISR) challenge organized in the Perception Beyond the Visible Spectrum (PBVS) 2022 workshop. The challenge uses the same thermal image dataset as the first two challenges, with 951 training images and 50 validation images at each resolution. A set of 20 images was kept aside for testing. The evaluation tasks were to measure the PSNR and SSIM between the SR image and the ground truth (HR thermal noisy image downsampled by four), and also to measure the PSNR and SSIM between the SR image and the semi-registered HR image (acquired with another camera). The results outperformed those from last year’s challenge, improving both evaluation metrics. This year, almost 100 teams participants registered for the challenge, showing the community’s interest in this hot topic.

1. Introduction

The goal of super-resolution in computer vision is to take a low-resolution image and turn it into a high-resolution image; most techniques used for this purpose are deep learning-based. These methods typically use a downsampled image from the high-resolution image as input, which is then augmented with noise and blur. The resulting image is then used to train the network. Most of these approaches have been used primarily in the visible spectrum, but with the increasing usage of thermal images for various applications, there is a need for methods that can operate in the thermal image domain.