Background The grooming process involves sexually explicit images or videos sent by the offender to the minor. Although offenders may try to conceal their identity, these sexts often include hand, knuckle, and nail bed imagery. Objective We present a novel biometric hand verification tool designed to identify online child sexual exploitation offenders from images or videos based on biometric/forensic features extracted from hand regions. The system can match and authenticate hand component imagery against a constrained custody suite reference of a known subject by employing advanced image processing and machine learning techniques. Data We conducted experiments on two hand datasets: Purdue University and Hong Kong. In particular, the Purdue dataset collected for this study allowed us to evaluate the system performance on various parameters, with specific emphasis on camera distance and orientation. Methods To explore the performance and reliability of the biometric verification models, we considered several parameters, including hand orientation, distance from the camera, single or multiple fingers, architecture of the models, and performance loss functions. Results Results showed the best performance for pictures sampled from the same database and with the same image capture conditions. Conclusion The authors conclude the biometric hand verification tool offers a robust solution that will operationally impact law enforcement by allowing agencies to investigate and identify online child sexual exploitation offenders more effectively. We highlight the strength of the system and the current limitations.

As a new research focus in biometrics, palm vein recognition has attracted people's attention because of high security, liveness-detection, user acceptability and convenience. Palm vein recognition exhibits high security, as it uses the network of blood vessels underneath palm skin for recognition. As palm vein is interior biological information of the body, vein patterns are much harder for intruders to copy compared to other biometric features. Palm vein is mostly invisible to the human eyes; they are commonly captured under near-infrared light. Under its natural state, the veins of the palm are concealed naturally and, for most of time people's hands are in the gesture of half fist. This study presents a comprehensive overview of recent research progress of palm vein recognition, from the basic background knowledge to data acquisition, public database, preprocessing, feature extraction and matching. In addition, the study focuses on palm vein-related fusion. Ultimately, the authors discuss the challenges and future perspectives in palm vein recognition for further works.

Palm vein recognition is motivated by the advantages of high security and liveness detection, but its popularity is prevented by the cost of palm vein capture devices. This study proposes a low-cost and practical palm vein recognition system. First, the authors' system captures near-infrared (NIR) palm vein image with complementary metal-oxide-semiconductor camera in lieu of an NIR charge-coupled device camera. The goal is to reduce the cost of palm vein capture devices greatly. Second, this study adopts thenar area on the palm as the region of interest (ROI) for further palm vein recognition. The goal is to get the rich vessel and avoid the effect of palmprint. Finally, the discriminate palm vein features are extracted based on Haar-wavelet decomposition and partial least squares algorithm on the ROI image. The goal is to increase the recognition accuracy, though the resolution of the image is low. A database with 1500 palm vein images from 250 samples is setup with the capture device. Experiments in the self-built database and a public database show the effectiveness of the scheme.

We present IJB–S dataset, an open-source IARPA Janus Surveillance Video Benchmark and associated protocols. The dataset consists of images and surveillance video collected from 202 subjects at a Department of Defense (DoD) training facility. Surveillance video was captured across multiple vignettes representative of a variety of real-world surveillance use cases that are particularly of interest to law enforcement and national security communities. Each video was annotated by human subject matter experts in order to generate ground truth identity and bounding box face labels. In total, over 10 million annotations were collected for the dataset. We present benchmark results utilizing state of the art deep learning approaches such as FaceNet. Our results illustrate and characterize the difficulty of the dataset.

In this paper, we develop a novel metric, which we call biometric permanence, to characterize the stability of biometric features. First, we define permanence in terms of the change in false non-match ratio (FNMR) over a repeated sequence of enrolment and verification events for a given population. We consider how such a measure may be experimentally determined. Since changes in FNMR, for most biometric modalities, are small, any variability in the biometric capture over time will camouflage the changes of interest. To address this issue, a robust methodology is proposed which can isolate the visit-to-visit variability, and substantially improve the estimation. We develop a model for the visit biases, and provide extensive simulation results supporting the efficacy of the improved method.

The purpose of this paper is to find whether there is any evidence of correlation between fingerprint quality and the factors of skin texture, keratin level, skin pigmentation, skin color, skin temperature, elasticity, and finger minutiae. In simpler terms, the goal was to see if and which finger characteristics affected the readability of the fingerprint. To achieve this goal, about 8000 random samples were collected from the fingers of 80 different subjects. The sensors collected data involving skin texture, keratin level, skin pigmentation, skin color, temperature, elasticity, and the amount of minutiae present on the finger. The sensors also collected the image quality of each fingerprint. This measurement is highly correlated with fingerprint scanner effectiveness and was therefore used as a representation of fingerprint readability in the experiment. A best subset test was run between the aforementioned factors and image quality in Minitab. This function tests all of the possible linear models that could be created by combining the factors against image quality and gives 2 results. The 1st result are the determined best models and the second are the statistics that tell the user how effective the models are. A model using all of the factors except pigmentation was used as the best model. However, this model only had an R2 value of 2.4, which meant that the model could only explains 2.4% of the image quality data. This provides strong evidence that there is no linear relationship between the factors and fingerprint image quality, and therefore fingerprint scanner effectiveness. In order to address the possibility of a nonlinear relationship between the factors and image quality, each factor was plotted on a graph against image quality. If the variable had a nonlinear relationship with image quality, a pattern would appear on the graph. No convincing pattern appeared on any of the graphs, which gave evidence that there is also no nonlinear relationship between the finger factors and image quality. This, combined with the previous finding concerning linear relationships, allows us to state that there is strong evidence that the factors do not correlate with fingerprint scanner effectiveness.

Biometric test reports are an important tool in the evaluation of biometric systems, and therefore the data entered into the system needs to be of the highest integrity. Data collection, especially across multiple modalities, can be a challenging experience for test administrators. They have to ensure that the data are collected properly, the test subjects are treated appropriately, and the test plan is followed. Tests become more complex as the number of sensors are increased, and therefore it becomes increasingly important that a test harness be developed to improve the accuracy of the data collection. This paper describes the development of a test harness for a complex multi-sensor, multi-visit data collection, and explains the processes for the development of such a harness. The applicability of such a software package for the broader biometric community is also considered.

Signature biometrics is a widely used form of user authentication. As a behavioural biometric, samples have inherent inconsistencies which must be accounted for within an automated system. Performance deterioration of a tuned biometric software system may be caused by an interaction error with a biometric capture device; however, using conventional error metrics, system and user interaction errors are combined, thereby masking the contribution by each element. In this paper we explore the application of the Human-Biometric Sensor Interaction (HBSI) model to signature as an exemplar of a behavioural biometric. Using observational data collected from a range of subjects, our study shows that usability issues can be identified specific to individual capture device technologies. While most interactions are successful, a range of common interaction errors need to be mitigated by design to reduce overall error rates.

Automated Border Control (ABC) in airports and land crossings utilize automated technology to verify passenger identity claims. Accuracy, interaction stability, user error, and the need for a harmonized approach to implementation are required. Two models proposed in this paper establish a global path through ABC processes. The first, the generic model, maps separately the enrolment and verification phases of an ABC scenario. This allows a standardization of the process and an exploration of variances and similarities between configurations across implementations. The second, the identity claim process, decomposes the verification phase of the generic model to an enhanced resolution of ABC implementations. Harnessing a human-biometric sensor interaction framework allows the identification and quantification of errors within the system's use, attributing these errors to either system performance or human interaction. Data from a live operational scenario are used to analyze behaviors, which aid in establishing what effect these have on system performance. Utilizing the proposed method will aid already established methods in improving the performance assessment of a system. Through analyzing interactions and possible behavioral scenarios from the live trial, it was observed that 30.96% of interactions included some major user error. Future development using our proposed framework will see technological advances for biometric systems that are able to categorize interaction errors and feedback appropriately.

Signatures are widely used as a form of personal authentication. Despite ubiquity in deployment, individual signatures are relatively easy to forge, especially when only the static 'pictorial' outcome of the signature is considered at verification time. In this study, we explore opinions on signature usage for verification purposes, and how individuals rate a particular third-party signature in terms of ease of forgeability and their own ability to forge. We examine responses with respect to an individual's experience of the forgeability/complexity of their own signature. Our study shows that past experience does not generally have an effect on perceived signature complexity nor the perceived effectiveness of an individual to themselves forge a signature. In assessing forgeability, most subjects cite the overall signature complexity and distinguishing features in reaching this decision. Furthermore, our research indicates that individuals typically vary their signature according to the scenario but generally little effort into the production of the signature.

This paper focused on the interoperability of nine different fingerprint sensors and examined the associated stability score index. The stability score index, conceptualized in 2013, was designed to address the weaknesses of the zoo menagerie and other performance metrics by quantifying the relative stability of a user from on condition to another. In this paper, the measure of interoperability was the stability score from enrolling on one sensor and verifying on multiple sensors. The results showed that, like the performance, individual performance was not stable across these sensors. When examining stability by sensor family (capacitance, optical and thermal), we find that capacitive as the enrollment sensor was the least stable. When individuals enrolled and verified on a thermal sensor, they were the most stable of the three family types. With respect to interaction type, enrolling on touch and verifying on swipe was more stable than enrolling on swipe and verifying on swipe, which was an interesting finding. Individuals who used the thermal sensor generated the most stable stability scores.

This research studied the question: “Are all individual’s performance stable in a fingerprint recognition system?” The fingerprints of 154 individuals, provided at different force levels, were examined using the biometric menagerie tool, first coined by Doddington et al. in 1998. The Biometric Menagerie illustrates how each person in a given dataset performs in a biometric system, by using their genuine and impostor scores, and providing them a classification based upon those scores. This research examined the biometric menagerie classifications across different force levels in a fingerprint recognition study to uncover if individuals performed the same over five force levels. The study concluded that they did not, and a new metric has been created to quantify this phenomenon. As a result of this discovery, the new metric, Stability Score Index is described to showcase the movement of individuals in the menagerie.

Mugshots taken by the Indiana Department of Corrections (IDOC) were analyzed for face recognition image quality. The mugshots provided by the IDOC was an assortment from different times and locations. 9,291 images were run through BSPALabs PreFace in the NIST Mugshot Best Practice Profile to observe compliance with ANSI/NIST-ITL 1-2000. Previous study done by Hale et al. revealed the noncompliance of IDOC images for NIST approved mugshots. The current test replicates Hale et al.'s methodology with latest update to the PreFace image quality analysis program in order to observe current compliance.

Biometric technologies represent a significant component of comprehensive digital identity solutions, and play an important role in crucial security tasks. These technologies support identification and authentication of individuals based on their physiological and behavioral characteristics. This has led many governmental agencies to choose biometrics as a supplement to existing identification schemes, most prominently ID cards and passports. Studies have shown that the success of biometric systems relies, in part, on how humans interact and accept such systems. In this paper, the authors build on previous work related to the Human-Biometric Sensor Interaction (HBSI) model and examine it with respect to the introduction of a token (e.g. an electronic passport or identity card) into the biometric system. The role of the imposter within an Identity Claim scenario has been integrated to expand the HBSI model into a full version, which is able to categorise potential False Claims and Attack Presentations.