Deep visible to thermal infrared style transfer in dynamic video sequences

Marc-André Blais; Moulay A. Akhloufi

doi:10.1117/12.2621608

27 May 2022 Deep visible to thermal infrared style transfer in dynamic video sequences

Proceedings Volume 12109, Thermosense: Thermal Infrared Applications XLIV; 121090I (2022) https://doi.org/10.1117/12.2621608
Event: SPIE Defense + Commercial Sensing, 2022, Orlando, Florida, United States

Abstract

Visible and thermal infrared are two imaging modalities which are used in a variety of applications. In deep learning we need large datasets to be able to train and optimize the algorithms. In thermal infrared imaging, there is a lack of large datasets. This work proposes a deep learning approach to transform visible light images into thermal infrared images using video sequences with moving objects. We propose to use and optimize a CycleGAN algorithm to transform frames from one spectrum to another by training two generators and two discriminators. The results are promising with impressive qualitative and quantitative results.

Conference Presentation

Citation Download Citation

Marc-André Blais and Moulay A. Akhloufi "Deep visible to thermal infrared style transfer in dynamic video sequences", Proc. SPIE 12109, Thermosense: Thermal Infrared Applications XLIV, 121090I (27 May 2022); https://doi.org/10.1117/12.2621608

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available