Abstract
Face recognition systems face real challenges from various presentation attacks. New, more sophisticated methods of presentation attacks are becoming more difficult to detect using traditional face recognition systems. Thermal infrared imaging offers specific physical properties that may boost presentation attack detection capabilities. The aim of this paper is to present outcomes of investigations on the detection of various face presentation attacks in thermal infrared in various conditions including thermal heating of masks and various states of subjects. A thorough analysis of presentation attacks using printed and displayed facial photographs, 3D-printed, custom flexible 3D-latex and silicone masks is provided. The paper presents the intensity analysis of thermal energy distribution for specific facial landmarks during long-lasting experiments. Thermalization impact, as well as varying the subject’s state due to physical effort on presentation attack detection are investigated. A new thermal face spoofing dataset is introduced. Finally, a two-step deep learning-based method for the detection of presentation attacks is presented. Validation results of a set of deep learning methods across various presentation attack instruments are presented.
Highlights
The presentation attack detection technology aims to determine whether the current subject is authentic
Results were almost equal across all presentation attack instruments (PAIs), the lowest scores were achieved for 2D paper mask attacks
Results were almost equal across all PAIs, the lowest scores were obtained for 2D paper mask attacks
Summary
The presentation attack detection technology aims to determine whether the current subject is authentic. This paper presents results of a study on presentation attack detection (PAD) in thermal infrared using simple two-dimensional attacks, as well as novel 3D facial masks. The study aims to analyze the impact of varying conditions on PAD performance. For the purpose of this study long lasting experiments have been performed to assess the impact of thermalization. A wide range of presentation attack instruments has been used including printed and displayed attacks, 2D paper-printed masks, 3D-printed masks, custom flexible latex full-face masks and custom silicone masks. Thermal images of subjects’ faces are collected in a long lasting scenario to evaluate thermalization impact of presentation attack detection, as well as individual heat patterns emitted by a human face are analyzed. In order to assess the impact of imagers’ parameters on PAD performance, studies are carried out using high-end and low-cost thermal imagers. A deep learning method for a non-intrusive PAD based on distribution of thermal energy emissions
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
