IET Biometrics最新文献

Signing a document or a cheque by hand or using a stored image-based signature is known to be a valid method for authentication and authorisation by the signer. However, signature forging has advanced to replicate exactly how a signature looks, which can be done by skilfully, unskilfully or randomly forging a signature. Such a dilemma presents a challenge to accurately authenticate and authorise using signatures. In this study, a verification system is proposed for handwritten image-based signatures for validating whether the image-based signature is authentic rather than forged. The system maps the live stream of an audio-based signature with the investigated image-based signature and returns the match results. Matching is done by classification and/or by correlation between the two signatures. If matching shows a similar class or a score above a pre-defined threshold, the image-based signature is verified to be authentic, otherwise it is flagged as forged. A total of 20 participated in the experiment, where each participant provided a legitimate signature and forged four other signatures in different settings. In a double-blind setting, the system reported 95% accuracy using a one-class SVM and 100% accuracy using a correlation coefficient for detecting forged versus legitimate signatures.

The intention of this position paper is to comment on the joint European Data Protection Supervisor (EDPS)-Agencia Española de Protección de Datos (aepd) publication ‘14 Misunderstandings with regard to Biometric Identification and Authentication’ that was published in June 2020 and to provide additional input to help with the better understanding of the issues raised in that publication. In particular, it aims to highlight some important missing information in the aforementioned publication. It is hoped that this paper will help with any future revision of the EDPS-aepd publication, such that it includes a full picture of the current state of the art in biometrics and the availability of standards and privacy enhancing techniques.

Recently, facial recognition has been extensively adopted in various fields. Wide applications are associated with a large amount of data transmission so that edge computing is inspired accordingly. In this research task, the major goal of edge computing is to handover a part of the computing work to the terminal equipment; the server only needs to process the results of final return. The IoT configuration proposed includes a perception layer, a transmission layer, and an application layer to fulfil a complete IoT system. In the perception layer, the facial authentication mechanism is adopted. This system is equipped with a highly robust anti-spoofing function, which can avoid criminal access from photos or electronic screens. Finally, the IoT transmission system is realised as the transmission layer. Combined with such a transmission mechanism, one can distribute user facial features to user's electronic devices instead of storing it in the server. This not only saves storage resources and transmission costs, but also allows users to complete data transmission and face authentication easily.

Electroencephalography (EEG)-based personal recognition in realistic contexts is still a matter of research, with the following issues to be clarified: (1) the duration of the signal length, called ‘epoch’, which must be very short for practical purposes and (2) the contribution of EEG sub-bands. These two aspects are connected because the shorter the epoch’s duration, the lower the contribution of the low-frequency sub-bands while enhancing the high-frequency sub-bands. However, it is well known that the former characterises the inner brain activity in resting or unconscious states. These sub-bands could be of no use in the wild, where the subject is conscious and not in the condition to put himself in a resting-state-like condition. Furthermore, the latter may concur much better in the process, characterising normal subject activity when awake. This study aims at clarifying the problems mentioned above by proposing a novel personal recognition architecture based on extremely short signal fragments called ‘patches’, subdividing each epoch. Patches are individually classified. A ‘qualified majority’ of classified patches allows taking the final decision. It is shown by experiments that this approach (1) can be adopted for practical purposes and (2) clarifies the sub-bands’ role in contexts still implemented in vitro but very similar to that conceivable in the wild.

Despite the impressive progress in the field of presentation attack detection and multimedia forensics over the last decade, these systems are still vulnerable to attacks in real-life settings. Some of the challenges for the existing solutions are the detection of unknown attacks, the ability to perform in adversarial settings, few-shot learning, and explainability. In this study, these limitations are approached by reliance on a game-theoretic view for modelling the interactions between the attacker and the detector. Consequently, a new optimisation criterion is proposed and a set of requirements are defined for improving the performance of these systems in real-life settings. Furthermore, a novel detection technique is proposed using generator-based feature sets that are not biased towards any specific attack species. To further optimise the performance on known attacks, a new loss function coined categorical margin maximisation loss (C-marmax) is proposed, which gradually improves the performance against the most powerful attack. The proposed approach provides a more balanced performance across known and unknown attacks and achieves state-of-the-art performance in known and unknown attack detection cases against rational attackers. Lastly, the few-shot learning potential of the proposed approach as well as its ability to provide pixel-level explainability is studied.

Inverse biometrics methods are a major privacy concern for users of biometric recognition systems. Affine-based reconstruction attack is an inverse biometrics method that models the biometric recognition algorithm by an affine approximation. This type of attack reconstructs targeted biometric references using the modelled biometric recognition algorithm and the comparison scores issued by the system. Although this reconstruction method has only been successfully applied to reconstruct face images, the common consensus is that any biometric system that issues comparison scores could be vulnerable to such an attack since this method is sufficiently general to be applied to other biometric templates. Here it is shown that the attack fails to regenerate sparse vascular feature point templates. The reconstruction attack on feature point patterns extracted from retina and hand vascular images is tested. The inverse attack match rate for reconstructed reference templates was 0.3% in one experiment using retinal vasculature and 0% for all others. These results show that the reconstruction attack is not as catastrophic as it is widely accepted to be, and that vascular biometric template protection schemes that store sparse templates as references and reveal comparison scores are not susceptible to affine-based reconstruction attacks.

Accurate segmentation and parameterisation of the iris in eye images still remain a significant challenge for achieving robust iris recognition, especially in off-angle images captured in less constrained environments. While deep learning techniques (i.e. segmentation-based convolutional neural networks (CNNs)) are increasingly being used to address this problem, there is a significant lack of information about the mechanism of the related distortions affecting the performance of these networks and no comprehensive recognition framework is dedicated, in particular, to off-angle iris recognition using such modules. In this work, the general effect of different gaze angles on ocular biometrics is discussed, and the findings are then related to the CNN-based off-angle iris segmentation results and the subsequent recognition performance. An improvement scheme is also introduced to compensate for some segmentation degradations caused by the off-angle distortions, and a new gaze-angle estimation and parameterisation module is further proposed to estimate and re-project (correct) the off-angle iris images back to frontal view. Taking benefit of these, several approaches (pipelines) are formulated to configure an end-to-end framework for the CNN-based off-angle iris segmentation and recognition. Within the framework of these approaches, a series of experiments is carried out to determine whether (i) improving the segmentation outputs and/or correcting the output iris images before or after the segmentation can compensate for some off-angle distortions, (ii) a CNN trained on frontal eye images is capable of detecting and extracting the learnt features on the corrected images, or (iii) the generalisation capability of the network can be improved by training it on iris images of different gaze angles. Finally, the recognition performance of the selected approach is compared against some state-of-the-art off-angle iris recognition algorithms.