Edge AI and Computer Vision Articles for 01.29.2021

Every day I take the time to look at the new computer vision articles on arXiv.org. This is my curated list of articles for January 29, 2021 related to either Edge AI or use cases relevant to industry.

Today we have one article which is a use case for detecting fires using a fusion of visible and infrared imagery.

Assessing the applicability of Deep Learning-based visible-infrared fusion methods for fire imagery

This paper examines fusing the two most common vision-based fire detection techniques: visible imagery and infrared imagery. The visible imagery solutions often suffer from false positives and false negatives caused by environmental conditions and illumination. Infrared imagery can overcome some of these challenges, but it is still not as accurate as required.

Three state-of-the-art and one new (proposed in the paper) infrared and visible imagery fusion approaches are applied to the fire detection problem evaluated and compared. And the results show that it is feasible to apply fusion to fire detection and that the new model FIRe-GAN shows the most promising results.

Source: Assessing the applicability of Deep Learning-based visible-infrared fusion methods for fire imagery
Source: Assessing the applicability of Deep Learning-based visible-infrared fusion methods for fire imagery

I hope that you find these articles as useful as I did. If you have additional thoughts, please add a comment.