Continuous authentication using biometrics: An advanced review

Abstract

The shortcomings of conventional access control systems for high‐security environments have led to the concert of continuous authentication. Contrary to traditional verification, in which users are authenticated only once at the start of their session, continuous authentication systems regularly check users' identities to prevent hijackings. The challenges in this area involve balancing the security of protected assets by quickly detecting intruders with the system usability for genuine users. Biometric recognition plays a major role within this context, as it is the main way to assure that users are who they claim to be. A comparative analysis of the latest works revealed different aspects of this problem. First, some biometrics traits among those applied for continuous authentication are more suitable for this task than others. Second, systems combining multiple traits have advantages over those relying on a single one. Finally, many works fail to report proper evaluation metrics. With this in mind, we were able to identify new opportunities for researchers in the field. We highlight the potential for mining new datasets on the internet, which would benefit validation and benchmarking, and how recent deep learning techniques could address some of the open challenges in the area. This article is categorized under: Technologies > Prediction Technologies > Machine Learning Application Areas > Science and Technology

Performance Optimization of Surface Electromyography Based Biometric Sensing System for Both Verification and Identification

“…3(C)). The EERr was very similar for all the methods for channel lengths (3)(4)(5)(6)(7)(8). For a combination of TD with spectral features it was observed that TD+FDT and TD+AR had similar values in all the three scenarios and for different channel lengths (p>0.05).…”

“…The TD method had a significantly lower R5E for 1-2 number of channels. However, for the inclusion of more channels (3)(4)(5)(6)(7)(8), there was no difference between the R5E for TD and FDT, but the AR had a significantly higher R5E than TD and FDT (p<0.05). For a combination of TD with spectral features it was observed that the R1E for TD+FDT was significantly lower (p<0.05) than the TD+AR and TD for different channel lengths.…”

“…Fig. 3 shows the distribution of the median EER and the median AUC across all participants, respectively for the three feature extraction methods (FDT, TD, and AR), their combinations (TD+FDT and TD+AR), and a different number of channels (1)(2)(3)(4)(5)(6)(7)(8). Median EER for the FDT method and for eight channels was 0.76% (Q1=0.1%, Q3=2.93%) for the Normal Test, 2.8% (Q1=0.3%, Q3=5.45%) for the Leaked Test, and 4.86% (Q1=2.1%, Q3=9.89%) for the Self Test (where Q1 is the 25 th percentile and Q3 is the 75 th percentile).…”

“…These biosignals are suggested to be used as an auxiliary biometric trait because they have two + https://dx.doi.org/10.21227/n9rt-p854 important disadvantages compared to well-developed well-established biometric traits. First of all, the performance accuracy is considerably lower [7]. Secondly, they are less robust against conditions such as illness, physical and mental activities, and the time interval between the training session and the real-time authentication session [8,9].…”

HOrBAC Optimization Based on Suspicious Behavior Detection Using Information Theory

“…This allowed us to redefine the emission process of the HOrBAC digital parafer by proposing a version with in behaviour detection. The proposed detection algorithm can be used for continuous authentication [1,11]. Our approach resulted in a reduction of 89.12% of the deferred requests issued according to the experiment.…”