Anti-spoofing in face recognition-based biometric authentication using Image Quality Assessment

Client-specific anomaly detection for face presentation attack detection

“…Although in any spoofing detection system the representations used are selected in a way that they capture the intrinsic differences between real-accesses and spoofing attempts, such an approach does not rule out the possibility that the features used can be affected by the specific characteristics of each client [8] . From this perspective, the majority of the work on face spoofing detection, including [3,6,9,10] , can be considered as clientindependent approaches. A client-independent face PAD approach assumes that the relevant information comes from either a realaccess or the attack class, whereas a client-specific method assumes that the constructed representations are additionally influenced by the identities of the subjects.…”

“…The comparison between the performance of anomaly detectors and two-class approaches using two different protocols in terms of frame-based HTER (%). 1.6 Patch-based CNN [9] 1.25 Depth-based CNN [9] 0.75 Fusion of the two Patch and Depth CNNs [9] 0.72 Deep discriminative feature maps [62] 0.3 Attention-based two-stream CNN [63] 0.25 Image Quality Assessment [10] 0.03 Deep Learning [64] 0 ( * ) Proposed Method 0 Replay-Mobile two-class SVM + Motion [6] 10.4 two-class SVM + Gabor [6] 9.13 Deep Pixel-wise [59] 0 ( * ) Proposed method 8.58 Rose-Youtu Wavelet [20] 26.6 CoALBP [20] 16.4 Deep Learning [20] 8.0 ( * ) Proposed method 8. 13 evaluation scheme considered in Chingovska [60] is unseen in the sense that it excludes one of the three attack types (Print, Digital Photo and Video) during training in each of the considered scenarios, the evaluation process in Edmunds and Caplier [15] cannot be considered completely unseen as the authors use similar attack types (video replays) both for training and evaluation in some of their evaluations.…”

A broad review on non-intrusive active user authentication in biometrics

“…Nonintrusive behavioural biometrics was utilized to provide security for the stored data in the computer system. It attempts to study the normal behaviour of authentic users on a system and then to identify abnormal behaviour from the normal behavioural pattern (Fourati et al 2020). Behavioural biometrics capture the characteristic patterns of the user's input, navigation through interfaces, and regular use patterns on both input devices as well as applications to create a virtual fingerprint of authentic user behaviour (Lang and Haar 2020).…”

Spoofing Detection on Hand Images Using Quality Assessment

“…The quality difference between the real and counterfeit images plays a key role for PAD. It has been studied that mainly the face Jia et al [2020], Pinto et al [2020], Fourati et al [2020], and finger-print Tolosana et al [2019] have received much research attention for PAD. In Jia et al [2020], PAD based on face images acquired using mobile devices is proposed.…”