Face detection in thermal imagery using an Open Source Computer Vision library

Sarun Sumriddetchkajorn; Armote Somboonkaew

doi:10.1117/12.819996

22 April 2009 Face detection in thermal imagery using an Open Source Computer Vision library

Sarun Sumriddetchkajorn, Armote Somboonkaew

Proceedings Volume 7299, Thermosense XXXI; 729906 (2009) https://doi.org/10.1117/12.819996
Event: SPIE Defense, Security, and Sensing, 2009, Orlando, Florida, United States

Abstract

This paper studies the use of a combination of Haar-like features and a cascade of boosted tree classifiers embedded in a widely used OpenCV for face detection in thermal images. With 2013 positive and 2020 negative 320×240-pixel thermal images for 20 training stages on three window sizes of 20×20, 24×24, and 30×30 pixels, our experiment shows that these three windows offer similar hit and false alarm rates at the end of the training section. Larger windows also spend much more time to train. During our testing, the 30×30-pixel window provides measured best hit and false rejection/acceptation rates of 93.4% and 6.6%, respectively, with a measured slowest detection speed of 19.6 ms. A 5-ms improvement in the measured detection speed with a slightly lower hit rate of 92.1% is accomplished by using the 24×24-pixel window. These results verify that the combination of Haar-like features and a cascade of boosted tree classifiers is a promising technique for face detection application in thermal images.

Citation Download Citation

Sarun Sumriddetchkajorn and Armote Somboonkaew "Face detection in thermal imagery using an Open Source Computer Vision library", Proc. SPIE 7299, Thermosense XXXI, 729906 (22 April 2009); https://doi.org/10.1117/12.819996

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