Vascular Biometric Graph Comparison: Theory and Performance

Abstract

Vascular biometric templates are gaining increasing popularity due to simple and contact free capture and resilience to presentation attacks. We present the state of the art in Biometric Graph Comparison, a technique to register and compare vascular biometric templates by representing them as formal graphs. Such graphs consist of a set of vertices, representing the branch, termination and crossover points in the vascular pattern, and a set of edges. An edge represents the relationship between a pair of feature points that are directly connected by a vessel segment in a vascular biometric image. We summarise how this information has been successfully used over the past 8 years to improve registration and recognition performance for the vasculature under the palm, wrist, hand and retina. The structural properties of biometric graphs from these modalities differ, with retina graphs having the largest number of vertices on average and the most complex structure, and hand graphs having the smallest number of vertices on average and being the least connected. All vascular graphs have similarities to trees, with the ratio of edges to vertices being close to 1. We describe our most recent algorithms for biometric graph registration and comparison, and our performance results. We are interested in the possibility of using biometric graphs in a template protection scheme based on the paradigm of dissimilarity vectors. As a first step, we wish to improve registration. Certain modalities like retina have an intrinsic reference frame that makes registration more straightforward. Other modalities may not have an intrinsic reference frame. To overcome this, we introduce the notion of anchors-subgraphs of a biometric graph, having between 5 and 10 vertices, that occur consistently in samples from the same individual-that would enable the dissimilarity vector scheme to be applied to any vascular modality. Experiments on palm and wrist databases show that all individuals had at least some sets of 6 captures which could be used to identify an anchor,

Failure of affine‐based reconstruction attack in regenerating vascular feature points

“…VARIA has 233 images of 139 individuals each having 768 � 584 pixels. The retinal point patterns are extracted from spatial graphs [18] that are rescaled and centred on the centre of the optic disc (OD). The registration sets the centre of OD at (0, 0) point.…”

“…The VARIA point pattern templates have fewer points on average than the ESRID templates and the SNIR point pattern templates have fewer points on average than the SFIR templates. The mean (standard deviation) of the number of feature points in ESRID and VARIA are 146.2 (86.7) and 70.3 (27.9), respectively, and for SFIR and SNIR are 51.7 (9.38) and 39.4 (13.07), respectively [18].…”

“…One application of this research is to quantify the security of the dissimilarity vector representation paradigm in [18] to test if it is a worthwhile candidate for developing a biometric template protection scheme when the templates are sparse. The cohort‐based biometric recognition system in [18] developed for vascular biometric graphs stores reference templates as distance vectors to the cohort.…”

“…One application of this research is to quantify the security of the dissimilarity vector representation paradigm in [18] to test if it is a worthwhile candidate for developing a biometric template protection scheme when the templates are sparse. The cohort‐based biometric recognition system in [18] developed for vascular biometric graphs stores reference templates as distance vectors to the cohort. In experiments performed in Section 4, we evaluate the risk of this attack on reconstructing point pattern templates assuming that the cohort templates are publicly known and the attacker has access to distance scores.…”

State of the Art in Vascular Biometrics

“…Finally, [9] rely on their approach of applying the Biometric Graph Matching (BGM) to graphs derived from skeletons of the vascular network. See a chapter in this book for a recent overview of this type of methodology [8].…”