Repozytorium publikacji - Politechnika Gdańska

Face detection is often a first step in quantitative analysis of face images. It is an important research area for visible images and recently also for thermography. Due to technological developments thermal cameras may be embedded into wearable devices to provide remote healthcare. In this paper, we compared three algorithms for face detection in thermal images by testing execution time, accuracy, symmetry ratio and false-positives. We investigated whether the computational performance of the methods allows for reliable face detection in real-time. The results show that the best algorithm allows to detect faces with acceptable accuracy (90.5±4.34%) in about 23ms per frame.

Autorzy

• Alicja Kwaśniewska link otwiera się w nowej karcie ,
• prof. dr hab. inż. Jacek Rumiński link otwiera się w nowej karcie