Thermal Image Super-Resolution
Overview
Thermal images, also known as thermograms, model radiation from the
infrared range of the electromagnetic spectrum. In many situations
(e.g., surveillance and reconnaissance, border and maritime patrol,
airborne law enforcement, search and rescue operations, critical
infrastructure protection, etc.) there exists a strong demand for
thermal imaging systems. However, due to underlying hardware
limitations, thermal imagery is inherently of low resolution. Thus,
there is an urgent need for solutions that allow for the ability to
regain lost image details. The goal of single image
super-resolution is to determine a mapping between a low-resolution
(LR) image and its high-resolution (HR) counterpart in the HR
space. Searching for this mapping is an ill-posed problem. Each LR
image has various mappings while the HR space is typically
intractable. Methods for finding this mapping are divided into
interpolated and learning-based, which can be represented by
bicubic interpolation and deep learning approaches, respectively.
Contributions
- We were finalists in the 2020 Thermal Image Super-Resolution Challenge
- We were finalists in the 2021 Thermal Image Super-Resolution Challenge
- We introduced a deep attention to varying receptive fields network to efficiently reconstruct thermal images
Publications
-
R.E. Rivadeneira, A.D. Sappa, B.X. Vintimilla, L. Guo, J. Hou, A. Merhi, P. Behjati, A.H. Patel, V. Chudasama, K. Prajapati, K.P. Upla, R. Ramachandra, K. Raja, C. Busch, F. Almasri, O. Debeir, S. Nathan, P. Kansal, N. Gutierrez, B. Mojra, and W.J. Beksi,
"Thermal Image Super-Resolution Challenge - PBVS 2020,"
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2020.
Paper • Citation -
R.E. Rivadeneira, A.D. Sappa, B.X. Vintimilla, S. Nathan, P. Kansal, A. Mehri, P.B. Ardakani, A. Dalal, A. Akula, D. Sharma, S. Pandey, B. Kumar, J. Yao, R. Wu, K. Feng, N. Li, Y. Zhao, H. Patel, V. Chudasama, K. Prajapati, A. Sarvaiya, K.P. Upla, K. Raja, R. Ramachandra, C. Busch, F. Almasri, T. Vandamme, O. Debeir, N.B. Gutierrez, Q.H. Nguyen, and W.J. Beksi,
"Thermal Image Super-Resolution Challenge - PBVS 2021,"
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2021.
Paper • Citation -
N.B. Gutierrez and W.J. Beksi,
"Thermal Image Super-Resolution Using Second-Order Channel Attention with Varying Receptive Fields,"
International Conference on Computer Vision Systems (ICVS), 2021.
Best Conference Paper Award Finalist • Paper • Preprint • Source Code • Citation