Thermal Features for Presentation Attack Detection in Hand Biometrics

Abstract

This paper proposes a method for utilizing thermal features of the hand for the purpose of presentation attack detection (PAD) that can be employed in a hand biometrics system's pipeline. By envisaging two different operational modes of our system, and by employing a DCNNbased classifiers fine-tuned with a dataset of real and fake hand representations captured in both visible and thermal spectrum, we were able to bring two important deliverables. First, a PAD method operating in an open-set mode, capable of correctly discerning 100% of fake thermal samples, achieving Attack Presentation Classification Error Rate (APCER) and Bona-Fide Presentation Classification Error Rate (BPCER) equal to 0%, which can be easily implemented into any existing system as a separate component. Second, a hand biometrics system operating in a closed-set mode, that has PAD built right into the recognition pipeline, and operating simultaneously with the userwise classification, achieving rank-1 recognition accuracy of up to 99.75%. We also show that thermal images of the human hand, in addition to liveness features they carry, can also improve classification accuracy of a biometric system, when coupled with visible light images. To follow the reproducibility guidelines and to stimulate further research in this area, we share the trained model weights, source codes, and a newly created dataset of fake hand representations with interested researchers.

Spoofing Detection on Hand Images Using Quality Assessment

“…On the other research line, though, handbased modalities such as fingerprint, palmprint, and finger-vein have been tested for PAD. However, to the best of our knowledge, only a few comprehensive works exist using hand geometry in the direction of spoofing attack detection Chen et al [2005], Bartuzi and Trokielewicz [2018]. An exemplary study on anti-spoofing, especially on hand-based modalities, is presented in Table 2.…”

“…However, no suitable experimental description has been illustrated. Bartuzi and Trokielewicz Bartuzi and Trokielewicz [2018] have proposed a method for PAD using thermal hand images. To create the fake hand dataset, the printed copy of the original hand images is used.…”

“…In Bartuzi and Trokielewicz [2018], thermal features of the hand are computed using a CNN architecture. It applies a score fusion of the RGB and thermal image matching modules to reach a final decision.…”

“…In the hardware-based method, an expensive sensor is employed for image acquisition from which the constraints about liveness such as the thermal facial map, blood pressure, perspiration, hand-vein temperature, and other salient features are computed Czajka and Bulwan [2013], Faundez-Zanuy et al [2014], Harvey et al [2018], Travieso et al [2014], Chen et al [2016]. Technological advancements of the sensor devices offer to acquire the thermal Czajka and Bulwan [2013], Faundez-Zanuy et al [2014], Travieso et al [2014], Chen et al [2016], Bartuzi and Trokielewicz [2018] and hyperspectral Ferrer et al [2014] images, which direct to devise high-security measures such as liveness detection to discard the fake hand biometric samples. On the contrary, the software-based method discriminates between the real and counterfeit images using a computable characteristic, such as visual quality assessment Zhang et al [2013], Reenu et al [2013], Sun et al [2018], Wang et al [2004], Gao et al [2018], Sellahewa and Jassim [2010], Guan et al [2017], Martini et al [2012].…”

“…Though hand geometry and other hand-based modalities are useful for identity verification in industry attendance maintenance, legal and forensic purposes Bera et al [2014], it is also exposed to direct attacks Chen et al [2005]. The spoofing attack detection (SAD) using the thermal hand image has been explored in Bartuzi and Trokielewicz [2018]. Individual authentication is not adequate to strengthen the security of a deployed hand biometric system.…”