We present the Membership Inference Test (MINT) Demo 2, a framework designed to improve transparency in machine learning training processes. MINT is a technique for experimentally determining whether specific data were used during machine learning model training. We establish the theoretical framework and propose multiple architectures for MINT depending on the amount of information known about the models that are being audited. Experimental results using a popular face recognition model, 4 state-of-the-art LLMs, and multiple, diverse, and large-scale public image and text databases achieve promising accuracy levels in the detection of training data of up to 90%. Building on these results, we introduce a comprehensive web platform1 that expands these capabilities to image and text modalities. The platform integrates a diverse technological stack, including MINT, aMINT, and gMINT, allowing users to audit a wide range of models. This demonstrator aims to promote AI transparency and provides a practical tool to foster compliance with emerging AI regulations.

Fatigue, sleep, or disturbances in daily activities are common symptoms among patients with neurodegenerative disorders (NDD) and immune-mediated inflammatory diseases (IMID). The current assessment of such symptoms is usually conducted using patient reported outcomes (PROs) based on standardized questionnaires that patients usually complete every few months. This assessment protocol has raised some concerns, due to its propensity to exhibit biases derived from its subjectivity nature, or the low sensibility to changes, which may lead to a failure when trying to capture variability over time. In this work, we explore the use of smartphone data, which can serve as a proxy for how patients interact with their devices, to provide an effective, reliable, and objective assessment of the symptoms mentioned above. Our study comprises data from 137 participants belonging to 6 different disease groups, plus a healthy control group. We conducted statistical analysis based on repeated measures correlation, in which we analyze the correlation between screen-time and app-usage features with scores obtained from the PROs collected from the participants using a smartphone application.

We present LoRA-MINT, a new methodology for Membership Inference Test (MINT) applied to recent Large Language Models (LLMs) fine-tuned for specific Natural Language Processing (NLP) tasks through Low-Rank Adaptation (LoRA). The primary goal is to assess whether individual samples were part of the training data of these adapted models, providing a useful auditing tool for the management of intellectual property and sensitive data. Our analysis explores the relationship between model perplexity and membership status, providing a systematic framework for estimating data exposure in fine-tuned LLMs. We conducted experiments on four models and three benchmark datasets, obtaining precision values in determining if given data were used for training ranging from 0.77 to 0.92, which outperform state-of-the-art baselines and demonstrate the robustness and generality of the proposed method. In general, our findings underscore the potential of LoRA-MINT as an effective and scalable framework for auditing LLMs, improving transparency, and fostering the ethical and responsible deployment of AI and NLP technologies. For the sake of concreteness and current relevance, our discussion and experiments are centered on LoRAadjusted LLMs, but note that most of the presented methodology is easily applicable for auditing training data given any other technique for adapting LLMs or, more generally, any other domain-adapted AI models.

• A new metric is proposed to compare statistical distributions, with a focus on detecting disparities in the distribution tails. • The method separately analyzes genuine and impostor score distributions to provide a view of bias in high-performance biometric systems. • The proposal was tested extensively on real-world, and large-scale operational scenarios. Demographic bias in high-performance face recognition (FR) systems often eludes detection by existing metrics, especially with respect to subtle disparities in the tails of the score distribution. We introduce the Comprehensive Equity Index (CEI), a novel metric designed to address this limitation. CEI uniquely analyzes genuine and impostor score distributions separately, enabling a configurable focus on tail probabilities while also considering overall distribution shapes. Our extensive experiments (evaluating state-of-the-art FR systems, intentionally biased models, and diverse datasets) confirm CEI’s superior ability to detect nuanced biases where previous methods fall short. Furthermore, we present CEI A , an automated version of the metric that enhances objectivity and simplifies practical application. CEI provides a robust and sensitive tool for operational FR fairness assessment. The proposed methods have been developed particularly for bias evaluation in face biometrics but, in general, they are applicable for comparing statistical distributions in any problem where one is interested in analyzing the distribution tails.

Human pose estimation has witnessed significant advancements in recent years, mainly due to the integration of deep learning models, the availability of a vast amount of data, and large computational resources. These developments have led to highly accurate body tracking systems, which have direct applications in sports analysis and performance evaluation. This work analyzes the performance of six trackers: two point trackers and four joint trackers for biomechanical analysis in sprints. The proposed framework compares the results obtained from these pose trackers with the manual annotations of biomechanical experts for more than 5870 frames. The experimental framework employs forty sprints from five professional runners, focusing on three key angles in sprint biomechanics: trunk inclination, hip flex extension, and knee flex extension. We propose a post-processing module for outlier detection and fusion prediction in the joint angles. The experimental results demonstrate that using joint-based models yields root mean squared errors ranging from 11.41° to 4.37°. When integrated with the post-processing modules, these errors can be reduced to 6.99° and 3.88°, respectively. The experimental findings suggest that human pose tracking approaches can be valuable resources for the biomechanical analysis of running. However, there is still room for improvement in applications where high accuracy is required.

Verifying the authenticity of identity documents (IDs) has become a critical challenge for real-life applications such as digital banking, crypto-exchanges, renting, etc. This study focuses on the topic of fake ID detection, covering several limitations in the field. In particular, there are no publicly available data from real IDs for proper research in this area, and most published studies rely on proprietary internal databases that are not available for privacy reasons. In order to advance this critical challenge of real data scarcity that makes it so difficult to advance the technology of machine learning-based fake ID detection, we introduce a new patch-based methodology that trades off privacy and performance, and propose a novel patch-wise approach for privacy-aware fake ID detection: FakeIDet. In our experiments, we explore: i) two levels of anonymization for an ID (i.e., fully- and pseudo-anonymized), and ii) different patch size configurations, varying the amount of sensitive data visible in the patch image. State-of-the-art methods, such as vision transformers and foundation models, are considered as backbones. Our results show that, on an unseen database (DLC-2021), our proposal for fake ID detection achieves 13.91% and 0% EERs at the patch and the whole ID level, showing a good generalization to other databases. In addition to the path-based methodology introduced and the new FakeIDet method based on it, another key contribution of our article is the release of the first publicly available database that contains 48,400 patches from real and fake IDs, called FakeIDet-db, together with the experimental framework1.

Person re-identification (ReID) across aerial and ground vantage points has become crucial for large-scale surveillance and public safety applications. Although significant progress has been made in ground-only scenarios, bridging the aerial-ground domain gap remains a formidable challenge due to extreme viewpoint differences, scale variations, and occlusions. Building upon the achievements of the AG-ReID 2023 Challenge, this paper introduces the AG-VPReID 2025 Challenge—the first large-scale video-based competition focused on high-altitude (80–120 m) aerial-ground person ReID. Constructed on the new AG-VPReID dataset with 3,027 identities, over 13,500 tracklets, and approximately 3.7 million frames captured from UAVs, CCTV, and wearable cameras, the challenge featured four international teams. These teams developed solutions ranging from multi-stream architectures to transformer-based temporal reasoning and physics-informed modeling. The leading approach, X-TFCLIP from UAM, attained 72.28% Rank-1 accuracy in the aerial-to-ground ReID setting and 70.77% in the ground-to-aerial ReID setting, surpassing existing baselines while highlighting the dataset’s complexity. For additional details, please refer to the official website at https://agvpreid25.github.io.

Photorealistic talking-head avatars are becoming increasingly common in virtual meetings, gaming, and social platforms. These avatars allow for more immersive communication, but they also introduce serious security risks. One emerging threat is impersonation: an attacker can steal a user’s avatar, preserving his appearance and voice,making it nearly impossible to detect its fraudulent usage by sight or sound alone. In this paper, we explore the challenge of biometric verification in such avatar-mediated scenarios. Our main question is whether an individual’s facial motion patterns can serve as reliable behavioral biometrics to verify their identity when the avatar’s visual appearance is a facsimile of its owner. To answer this question, we introduce a new dataset of realistic avatar videos created using a stateof-the-art one-shot avatar generation model, GAGAvatar, with genuine and impostor avatar videos. We also propose a lightweight, explainable spatio-temporal Graph Convolutional Network architecture with temporal attention pooling, that uses only facial landmarks to model dynamic facial gestures. Experimental results demonstrate that facial motion cues enable meaningful identity verification with AUC values approaching 80%. The proposed benchmark and biometric system are available<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> for the research community in order to bring attention to the urgent need for more advanced behavioral biometric defenses in avatar-based communication systems.

Behavioral biometrics based on smartphone motion sensors are growing in popularity for authentication purposes. In this study, AirSignatureDB is presented: a new publicly accessible dataset of in-air signatures collected from 108 participants under real-world conditions, using 83 different smartphone models across four sessions. This dataset includes genuine samples and skilled forgeries, enabling a comprehensive evaluation of system robustness against realistic attack scenarios. Traditional and deep learningbased methods for in-air signature verification are benchmarked, while analyzing the influence of sensor modality and enrollment strategies. Beyond verification, a first approach to reconstructing the three-dimensional trajectory of in-air signatures from inertial sensor data alone is introduced. Using on-line handwritten signatures as a reference, we demonstrate that the recovery of accurate trajectories is feasible, challenging the long-held assumption that in-air gestures are inherently traceless. Although this approach enables forensic traceability, it also raises critical questions about the privacy boundaries of behavioral biometrics. Our findings underscore the need for a reevaluation of the privacy assumptions surrounding inertial sensor data, as they can reveal user-specific information that had not previously been considered in the design of in-air signature systems.

Active Membership Inference Test (aMINT) is a method designed to detect whether given data were used during the training of machine learning models. In Active MINT, we propose a novel multitask learning process that involves training simultaneously two models: the original or Audited Model, and a secondary model, referred to as the MINT Model, responsible for identifying the data used for training the Audited Model. This novel multi-task learning approach has been designed to incorporate the auditability of the model as an optimization objective during the training process of neural networks. The proposed approach incorporates intermediate activation maps as inputs to the MINT layers, which are trained to enhance the detection of training data. We present results using a wide range of neural networks, from lighter architectures such as MobileNet to more complex ones such as Vision Transformers, evaluated in 5 public benchmarks. Our proposed Active MINT achieves over 80% accuracy in detecting if given data was used for training, significantly outperforming previous approaches in the literature. Our aMINT and related methodological developments contribute to increasing transparency in AI models, facilitating stronger safeguards in AI deployments to achieve proper security, privacy, and copyright protection.

Abstract Maintaining a healthy lifestyle has become increasingly challenging in today’s sedentary society marked by poor eating habits. To address this issue, both national and international organisations have made numerous efforts to promote healthier diets and increased physical activity. However, implementing these recommendations in daily life can be difficult, as they are often generic and not tailored to individuals. This study presents the AI4Food-NutritionDB database, the first nutrition database that incorporates food images and a nutrition taxonomy based on recommendations by national and international health authorities. The database offers a multi-level categorisation, comprising 6 nutritional levels, 19 main categories (e.g., “Meat”), 73 subcategories (e.g., “White Meat”), and 893 specific food products (e.g., “Chicken”). The AI4Food-NutritionDB opens the doors to new food computing approaches in terms of food intake frequency, quality, and categorisation. Also, we present a standardised experimental protocol and benchmark including three tasks based on the nutrition taxonomy (i.e., category, subcategory, and final product recognition). These resources are available to the research community, including our deep learning models trained on AI4Food-NutritionDB, which can serve as pre-trained models, achieving accurate recognition results for challenging food image databases. All these resources are available in GitHub ( https://github.com/BiDAlab/AI4Food-NutritionDB ).

This article presents the Keystroke Verification Challenge - onGoing (KVC-onGoing) 1 1 https://sites.google.com/view/bida-kvc/ . , on which researchers can easily benchmark their systems in a common platform using large-scale public databases, the Aalto University Keystroke databases, and a standard experimental protocol. The keystroke data consist of tweet-long sequences of variable transcript text from over 185,000 subjects, acquired through desktop and mobile keyboards simulating real-life conditions. The results on the evaluation set of KVC-onGoing have proved the high discriminative power of keystroke dynamics, reaching values as low as 3.33% of Equal Error Rate (EER) and 11.96% of False Non-Match Rate (FNMR) @1% False Match Rate (FMR) in the desktop scenario, and 3.61% of EER and 17.44% of FNMR @1% at FMR in the mobile scenario, significantly improving previous state-of-the-art results. Concerning demographic fairness, the analyzed scores reflect the subjects’ age and gender to various extents, not negligible in a few cases. The framework runs on CodaLab 2 2 https://codalab.lisn.upsaclay.fr/competitions/14063 . . • We set up a novel framework for developing and evaluating keystroke biometrics. • We designed a unified experimental protocol with desktop and mobile scenarios. • We employ the biggest databases of keystroke dynamics, with over 185,000 subjects. • We provide a competitive performance baseline based on a limited-time challenge. • We provide a first exploration of the biometric fairness of keystroke dynamics.

In 2021, the pioneering work TypeNet showed that keystroke dynamics verification could scale to hundreds of thousands of users with minimal performance degradation. Recently, the KVC-onGoing competition1 has provided an open and robust experimental protocol for evaluating keystroke dynamics verification systems of such scale. This article describes Type2Branch, the model and techniques that achieved the lowest error rates at the KVC-onGoing, in both desktop and mobile typing scenarios. The novelty aspects of the proposed Type2Branch include:i) synthesized timing features emphasizing user behavior deviation from the general population,ii) a dual-branch architecture combining recurrent and convolutional paths with various attention mechanisms,iii) a new loss function named Set2set that captures the global structure of the embedding space, andiv) a training curriculum of increasing difficulty. Considering five enrollment samples per subject of approximately 50 characters typed, the proposed Type2Branch achieves state-of-the-art performance with mean per-subject Equal Error Rates (EERs) of 0.77% and 1.03% on evaluation sets of respectively 15,000 and 5,000 subjects for desktop and mobile scenarios. With a fixed global threshold for all subjects, the EERs are respectively 3.25% and 3.61% for desktop and mobile scenarios, outperforming previous approaches by a significant margin. The source code for dataset generation, model, and training process is publicly available2.

Large-scale face recognition datasets are collected by crawling the Internet and without individuals' consent, raising legal, ethical, and privacy concerns. With the recent advances in generative models, recently several works proposed generating synthetic face recognition datasets to mitigate concerns in web-crawled face recognition datasets. This paper presents the summary of the Synthetic Data for Face Recognition (SDFR) Competition held in conjunction with the 18th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2024) and established to investigate the use of synthetic data for training face recognition models. The SDFR competition was split into two tasks, allowing participants to train face recognition systems using new synthetic datasets and/or existing ones. In the first task, the face recognition backbone was fixed and the dataset size was limited, while the second task provided almost complete freedom on the model backbone, the dataset, and the training pipeline. The submitted models were trained on existing and also new synthetic datasets and used clever methods to improve training with synthetic data. The submissions were evaluated and ranked on a diverse set of seven benchmarking datasets. The paper gives an overview of the submitted face recognition models and reports achieved performance compared to baseline models trained on real and synthetic datasets. Furthermore, the evaluation of submissions is extended to bias assessment across different demography groups. Lastly, an outlook on the current state of the research in training face recognition models using synthetic data is presented, and existing problems as well as potential future directions are also discussed.

The growing number of mobile devices over the past few years brings a large amount of personal information, which needs to be properly protected. As a result, several mobile authentication methods have been developed. In particular, behavioural biometrics has become one of the most relevant methods due to its ability to extract the uniqueness of each subject in a secure, non-intrusive, and continuous way. This article presents SwipeFormer, a novel Transformer-based system for mobile subject authentication by means of swipe gestures in an unconstrained scenario (i.e., subjects could use their personal devices freely, without restrictions on the direction of swipe gestures or the position of the device). Our proposed system contains two modules: (i) a Transformer-based feature extractor, and (ii) a similarity computation module. Mobile data from the touchscreen and different background sensors (accelerometer and gyroscope) have been studied, including in the analysis both Android and iOS operating systems. A complete analysis of SwipeFormer is carried out using an in-house large-scale database acquired in unconstrained scenarios. In these operational conditions, SwipeFormer achieves Equal Error Rate (EER) values of 6.6% and 3.6% on Android and iOS respectively, outperforming the state of the art. In addition, we evaluate SwipeFormer on the popular publicly available databases Frank DB and HuMIdb, achieving EER values of 11.0% and 5.0% respectively, outperforming previous approaches under the same experimental setup.

Abstract This article proposes a novel children–computer interaction (CCI) approach for the task of age group detection. This approach focuses on the automatic analysis of the time series generated from the interaction of the children with mobile devices. In particular, we extract a set of 25 time series related to spatial, pressure, and kinematic information of the children interaction while colouring a tree through a pen stylus tablet, a specific test from the large-scale public ChildCIdb database. A complete analysis of the proposed approach is carried out using different time series selection techniques to choose the most discriminative ones for the age group detection task: (i) a statistical analysis and (ii) an automatic algorithm called sequential forward search (SFS). In addition, different classification algorithms such as dynamic time warping barycenter averaging (DBA) and hidden Markov models (HMM) are studied. Accuracy results over 85% are achieved, outperforming previous approaches in the literature and in more challenging age group conditions. Finally, the approach presented in this study can benefit many children-related applications, for example, towards an age-appropriate environment with the technology.

Abstract The broad usage of mobile devices nowadays, the sensitiveness of the information contained in them, and the shortcomings of current mobile user authentication methods are calling for novel, secure, and unobtrusive solutions to verify the users’ identity. In this article, we propose TypeFormer, a novel transformer architecture to model free-text keystroke dynamics performed on mobile devices for the purpose of user authentication. The proposed model consists in temporal and channel modules enclosing two long short-term memory recurrent layers, Gaussian range encoding, a multi-head self-attention mechanism, and a block-recurrent transformer layer. Experimenting on one of the largest public databases to date, the Aalto mobile keystroke database, TypeFormer outperforms current state-of-the-art systems achieving equal error rate values of 3.25% using only five enrolment sessions of 50 keystrokes each. In such way, we contribute to reducing the traditional performance gap of the challenging mobile free-text scenario with respect to its desktop and fixed-text counterparts. To highlight the design rationale, an analysis of the experimental results of the different modules implemented in the development of TypeFormer is carried out. Additionally, we analyse the behaviour of the model with different experimental configurations such as the length of the keystroke sequences and the amount of enrolment sessions, showing margin for improvement.

Abstract International Organizations urge the protection of our oceans and their ecosystems due to their immeasurable importance to humankind. Since illegal fishing activities, commonly known as IUU fishing, cause irreparable damage to these ecosystems, concerned organisms are pushing to detect and combat IUU fishing practices. The automatic identification system allows to locate the position and trajectory of fishing vessels. In this study we address the task of detecting vessels’ fishing gears based on the trajectory behavior defined by GPS position data, a useful task to prevent the proliferation of IUU fishing practices. We present a new database including trajectories that span 7 different fishing gears and analyze these as in a time sequence analysis problem. We leverage from feature extraction techniques from the online signature verification domain to model vessel trajectories, and extract relevant information in the form of both local and global feature sets. We show how, based on these sets of features, the kinematics of vessels according to different fishing gears can be effectively classified using common supervised learning algorithms with accuracies up to $$90\%$$ 90 % . Furthermore, motivated by the concerns raised by several organizations on the adverse impact of bottom trawling on marine biodiversity, we present a binary classification experiment in which we were able to distinguish this kind of fishing gear with an accuracy of $$99\%$$ 99 % . We also illustrate in an ablation study the relevance of factors such as data availability and the sampling period to perform fishing gear classification. Compared to existing works, we highlight these factors, especially the importance of using sampling periods in the order of minutes instead of hours.

This work introduces a new multispectral database and framework to train and evaluate eyeblink detection in RGB and Near-Infrared (NIR). Our contributed dataset (mEBAL2, multimodal EyeBlink and Attention Level estimation, Version 2) is the largest existing eyeblink database, representing a great opportunity to improve data-driven multispectral approaches for blink detection and related applications (e.g., attention level estimation). mEBAL2 includes 21,100 image sequences from 180 different students (more than 2 million labeled images in total) while conducting a number of e-learning tasks of varying difficulty or taking a real course on HTML initiation through the edX MOOC platform. mEBAL2 uses multiple sensors, including two Near-Infrared (NIR) and one RGB camera to capture facial gestures during the execution of the tasks, as well as an Electroencephalogram (EEG) band to get the cognitive activity of the user and blinking events. Furthermore, this work proposes 3 data-driven approaches as benchmarks for blink detection on mEBAL2, where the architecture based on Convolutional Long Short-Term Memory (ConvLSTM) achieved performances of up to 99%. The experiments explored whether combining RGB and NIR spectrum data improves blink detection in training and architectures that merge both types of data. Experiments showed that the NIR spectrum enhances results, even when only RGB images are available during inference. Finally, the generalization capacity of the proposed eyeblink detectors, along with state-of-the-art eyeblink detection implementations, is validated in wilder and more challenging environments like the HUST-LEBW dataset to show the usefulness of mEBAL2 to train a new generation of data-driven approaches for eyeblink detection.

We present edBB-Demo, a demonstrator of an AI-powered research platform for student monitoring in remote education. The edBB platform aims to study the challenges associated to user recognition and behavior understanding in digital platforms. This platform has been developed for data collection, acquiring signals from a variety of sensors including keyboard, mouse, webcam, microphone, smartwatch, and an Electroencephalography band. The information captured from the sensors during the student sessions is modelled in a multimodal learning framework. The demonstrator includes: i) Biometric user authentication in an unsupervised environment; ii) Human action recognition based on remote video analysis; iii) Heart rate estimation from webcam video; and iv) Attention level estimation from facial expression analysis.

Institute, Warsaw, Poland This paper describes the results of the IEEE BigData 2023 Keystroke Verification Challenge1(KVC), that considers the biometric verification performance of Keystroke Dynamics (KD), captured as tweet-long sequences of variable transcript text from over 185,000 subjects. The data are obtained from two of the largest public databases of KD up to date, the Aalto Desktop and Mobile Keystroke Databases, guaranteeing a minimum amount of data per subject, age and gender annotations, absence of corrupted data, and avoiding excessively unbalanced subject distributions with respect to the considered demographic attributes. Several neural architectures were proposed by the participants, leading to global Equal Error Rates (EERs) as low as 3.33% and 3.61% achieved by the best team respectively in the desktop and mobile scenario, outperforming the current state of the art biometric verification performance for KD. Hosted on CodaLab2, the KVC will be made ongoing to represent a useful tool for the research community to compare different approaches under the same experimental conditions and to deepen the knowledge of the field.

Presentation Attack Detection (PAD) is a crucial stage in facial recognition systems to avoid leakage of personal information or spoofing of identity to entities. Recently, pulse detection based on remote photoplethysmography (rPPG) has been shown to be effective in face presentation attack detection. This work presents three different approaches to the presentation attack detection based on rPPG: (i) The physiological domain, a domain using rPPG-based models, (ii) the Deepfakes domain, a domain where models were retrained from the physiological domain to specific Deepfakes detection tasks; and (iii) a new Presentation Attack domain was trained by applying transfer learning from the two previous domains to improve the capability to differentiate between bona-fides and attacks. The results show the efficiency of the rPPG-based models for presentation attack detection, evidencing a 21.70% decrease in average classification error rate (ACER) (from 41.03% to 19.32%) when the presentation attack domain is compared to the physiological and Deepfakes domains. Our experiments highlight the efficiency of transfer learning in rPPG-based models and perform well in presentation attack detection in instruments that do not allow copying of this physiological feature.

Among user authentication methods, behavioural biometrics has proven to be effective against identity theft as well as user-friendly and unobtrusive. One of the most popular traits in the literature is keystroke dynamics due to the large deployment of computers and mobile devices in our society. This paper focuses on improving keystroke biometric systems on the free-text scenario. This scenario is characterised as very challenging due to the uncontrolled text conditions, the influence of the user's emotional and physical state, and the in-use application. To overcome these drawbacks, methods based on deep learning such as Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) have been proposed in the literature, outperforming traditional machine learning methods. However, these architectures still have aspects that need to be reviewed and improved. To the best of our knowl-edge, this is the first study that proposes keystroke biometric systems based on Transformers. The proposed Transformer architecture has achieved Equal Error Rate (EER) values of 3.84% in the popular Aalto mobile keystroke database using only 5 enrolment sessions, outperforming by a large margin other state-of-the-art approaches in the literature.

This study investigates the possibility of mitigating the demographic biases that affect face recognition technologies through the use of synthetic data. Demographic biases have the potential to impact individuals from specific demographic groups, and can be identified by observing disparate performance of face recognition systems across demographic groups. They primarily arise from the unequal representations of demographic groups in the training data. In recent times, synthetic data have emerged as a solution to some problems that affect face recognition systems. In particular, during the generation process it is possible to specify the desired demographic and facial attributes of images, in order to control the demographic distribution of the synthesized dataset, and fairly represent the different demographic groups. We propose to fine-tune with synthetic data existing face recognition systems that present some demographic biases. We use synthetic datasets generated with GANDiffFace, a novel framework able to synthesize datasets for face recognition with controllable demographic distribution and realistic intra-class variations. We consider multiple datasets representing different demographic groups for training and evaluation. Also, we fine-tune different face recognition systems, and evaluate their demographic fairness with different metrics. Our results support the proposed approach and the use of synthetic data to mitigate demographic biases in face recognition.

Mobile devices such as smartphones and smartwatches are part of our everyday life, acquiring large amount of personal information that needs to be properly secured. Among the different authentication techniques, behavioural biometrics has become a very popular method as it allows authentication in a non-intrusive and continuous way. This study proposes M-GaitFormer, a novel mobile biometric gait verification system based on Transformer architectures. This biometric system only considers the accelerometer and gyroscope data acquired by the mobile device. A complete analysis of the proposed M-GaitFormer is carried out using the popular available databases whuGAIT and OU-ISIR. M-GaitFormer achieves Equal Error Rate (EER) values of 3.42% and 2.90% on whuGAIT and OU-ISIR, respectively, outperforming other state-of-the-art approaches based on popular Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs).

Biometrics on mobile devices has attracted a lot of attention in recent years as it is considered a user-friendly authentication method. This interest has also been motivated by the success of Deep Learning (DL). Architectures based on Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) have established convenience for the task, improving the performance and robustness in comparison to traditional machine learning techniques. However, some aspects must still be revisited and improved. To the best of our knowledge, this is the first article that explores and proposes a novel gait biometric recognition systems based on Transformers, which currently obtain state-of-the-art performance in many applications. Several state-of-the-art architectures (Vanilla, Informer, Autoformer, Block-Recurrent Transformer, and THAT) are considered in the experimental framework. In addition, new Transformer configurations are proposed to further increase the performance. Experiments are carried out using the two popular public databases: whuGAIT and OU-ISIR. The results achieved prove the high ability of the proposed Transformer, outperforming state-of-the-art CNN and RNN architectures.

Mobile behavioral biometrics have become a popular topic of research, reaching promising results in terms of authentication, exploiting a multimodal combination of touchscreen and background sensor data. However, there is no way of knowing whether state-of-the-art classifiers in the literature can distinguish between the notion of user and device. In this article, we present a new database, BehavePassDB, structured into separate acquisition sessions and tasks to mimic the most common aspects of mobile Human-Computer Interaction (HCI). BehavePassDB is acquired through a dedicated mobile app installed on the subjects devices, also including the case of different users on the same device for evaluation. We propose a standard experimental protocol and benchmark for the research community to perform a fair comparison of novel approaches with the state of the art1. We propose and evaluate a system based on Long-Short Term Memory (LSTM) architecture with triplet loss and modality fusion at score level.

This paper describes the experimental framework and results of the IJCB 2022 Mobile Behavioral Biometrics Competition (MobileB2C). The aim of MobileB2C is bench-marking mobile user authentication systems based on behavioral biometric traits transparently acquired by mobile devices during ordinary Human-Computer Interaction (HCI), using a novel public database, BehavePassDB11https://github.com/BiDAlab/MobileB2C_BehavePassDE, and a standard experimental protocol. The competition is divided into four tasks corresponding to typical user activities: keystroke, text reading, gallery swiping, and tapping. The data are composed of touchscreen data and several background sensor data simultaneously acquired. “Random” (different users with different devices) and “skilled” (different user on the same device attempting to imitate the legitimate one) impostor scenarios are considered. The results achieved by the participants show the feasibility of user authentication through behavioral biometrics, although this proves to be a non-trivial challenge. MobileB2C will be established as an on-going competition22https://sites.google.com/view/mobileb2c/.

Media forensics has attracted a tremendous attention in the last years in part due to the increasing concerns around DeepFakes. Since the release of the initial DeepFakes databases of the 1st generation such as UADFV and FaceForensics++ up to the latest databases of the 2nd generation such as Celeb-DF and DFDC, many visual improvements have been carried out, making fake videos almost indistinguishable to the human eye. This study provides an in-depth analysis of both 1st and 2nd DeepFakes generations in terms of fake detection performance. Two different methods are considered in our experimental framework: (i) the traditional one followed in the literature based on selecting the entire face as input to the fake detection system, and (ii) a novel approach based on the selection of specific facial regions as input to the fake detection system. Fusion techniques are applied both to the facial regions and also to three different state-of-the-art fake detection systems (Xception, Capsule Network, and DSP-FWA) in order to further increase the robustness of the detectors considered. Finally, experiments regarding intra- and inter-database scenarios are performed. Among all the findings resulting from our experiments, we highlight: (i) the very good results achieved using facial regions and fusion techniques with fake detection results above 99% Area Under the Curve (AUC) for UADFV, FaceForensics++, and Celeb-DF v2 databases, and (ii) the necessity to put more efforts on the analysis of inter-database scenarios to improve the ability of the fake detectors against attacks unseen during learning.

This work presents a new deep learning approach for keystroke biometrics based on a novel Distance Metric Learning method (DML). DML maps input data into a learned representation space that reveals a “semantic” structure based on distances. In this work, we propose a novel DML method specifically designed to address the challenges associated to free-text keystroke identification where the classes used in learning and inference are disjoint. The proposed SetMargin Loss (SM-L) extends traditional DML approaches with a learning process guided by pairs of sets instead of pairs of samples, as done traditionally. The proposed learning strategy allows to enlarge inter-class distances while maintaining the intra-class structure of keystroke dynamics. We analyze the resulting representation space using the mathematical problem known as Circle Packing, which provides neighbourhood structures with a theoretical maximum inter-class distance. We finally prove experimentally the effectiveness of the proposed approach on a challenging task: keystroke biometric identification over a large set of 78,000 subjects. Our method achieves state-of-the-art accuracy on a comparison performed with the best existing approaches.

This article presents SVC-onGoing1, an on-going competition for on-line signature verification where researchers can easily benchmark their systems against the state of the art in an open common platform using large-scale public databases, such as DeepSignDB2 and SVC2021_EvalDB3, and standard experimental protocols. SVC-onGoing is based on the ICDAR 2021 Competition on On-Line Signature Verification (SVC 2021), which has been extended to allow participants anytime. The goal of SVC-onGoing is to evaluate the limits of on-line signature verification systems on popular scenarios (office/mobile) and writing inputs (stylus/finger) through large-scale public databases. Three different tasks are considered in the competition, simulating realistic scenarios as both random and skilled forgeries are simultaneously considered on each task. The results obtained in SVC-onGoing prove the high potential of deep learning methods in comparison with traditional methods. In particular, the best signature verification system has obtained Equal Error Rate (EER) values of 3.33% (Task 1), 7.41% (Task 2), and 6.04% (Task 3). Future studies in the field should be oriented to improve the performance of signature verification systems on the challenging mobile scenarios of SVC-onGoing in which several mobile devices and the finger are used during the signature acquisition.

Numerous studies in the literature have already shown the potential of biometrics on mobile devices for authentication purposes. However, it has been shown that, the learning processes associated to biometric systems might expose sensitive personal information about the subjects. This study proposes GaitPrivacyON, a novel mobile gait biometrics verification approach that provides accurate authentication results while preserving the sensitive information of the subject. It comprises two modules: i) two convolutional Autoencoders with shared weights that transform attributes of the biometric raw data, such as the gender or the activity being performed, into a new privacy-preserving representation; and ii) a mobile gait verification system based on the combination of Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) with a Siamese architecture. The main advantage of GaitPrivacyON is that the first module (convolutional Autoencoders) is trained in an unsupervised way, without specifying the sensitive attributes of the subject to protect. Two experimental studies have been examinated: i) MotionSense and MobiAct databases; and ii) OU-ISIR database. The experimental results achieved suggest the potential of GaitPrivacyON to significantly improve the privacy of the subject while keeping user authentication results higher than 96.6% Area Under the Curve (AUC). To the best of our knowledge, this is the first mobile gait verification approach that considers privacy-preserving methods trained in an unsupervised way.

Current mobile user authentication systems based on PIN codes, fingerprint, and face recognition have several shortcomings. Such limitations have been addressed in the literature by exploring the feasibility of passive authentication on mobile devices through behavioral biometrics. In this line of research, this work carries out a comparative analysis of unimodal and multimodal behavioral biometric traits acquired while the subjects perform different activities on the phone such as typing, scrolling, drawing a number, and tapping on the screen, considering the touchscreen and the simultaneous background sensor data (accelerometer, gravity sensor, gyroscope, linear accelerometer, and magnetometer). Our experiments are performed over HuMIdb,1 one of the largest and most comprehensive freely available mobile user interaction databases to date. A separate Recurrent Neural Network (RNN) with triplet loss is implemented for each single modality. Then, the weighted fusion of the different modalities is carried out at score level. In our experiments, the most discriminative background sensor is the magnetometer, whereas among touch tasks the best results are achieved with keystroke in a fixed-text scenario. In all cases, the fusion of modalities is very beneficial, leading to Equal Error Rates (EER) ranging from 4% to 9% depending on the modality combination in a 3-second interval.

This study proposes DeepWriteSYN, a novel on-line handwriting synthesis approach via deep short-term representations. It comprises two modules: i) an optional and interchangeable temporal segmentation, which divides the handwriting into short-time segments consisting of individual or multiple concatenated strokes; and ii) the on-line synthesis of those short-time handwriting segments, which is based on a sequence-to-sequence Variational Autoencoder (VAE). The main advantages of the proposed approach are that the synthesis is carried out in short-time segments (that can run from a character fraction to full characters) and that the VAE can be trained on a configurable handwriting dataset. These two properties give a lot of flexibility to our synthesiser, e.g., as shown in our experiments, DeepWriteSYN can generate realistic handwriting variations of a given handwritten structure corresponding to the natural variation within a given population or a given subject. These two cases are developed experimentally for individual digits and handwriting signatures, respectively, achieving in both cases remarkable results. Also, we provide experimental results for the task of on-line signature verification showing the high potential of DeepWriteSYN to improve significantly one-shot learning scenarios. To the best of our knowledge, this is the first synthesis approach capable of generating realistic on-line handwriting in the short term (including handwritten signatures) via deep learning. This can be very useful as a module toward long-term realistic handwriting generation either completely synthetic or as natural variation of given handwriting samples.

This work proposes a novel privacy-preserving neural network feature representation to suppress the sensitive information of a learned space while maintaining the utility of the data. The new international regulation for personal data protection forces data controllers to guarantee privacy and avoid discriminative hazards while managing sensitive data of users. In our approach, privacy and discrimination are related to each other. Instead of existing approaches aimed directly at fairness improvement, the proposed feature representation enforces the privacy of selected attributes. This way fairness is not the objective, but the result of a privacy-preserving learning method. This approach guarantees that sensitive information cannot be exploited by any agent who process the output of the model, ensuring both privacy and equality of opportunity. Our method is based on an adversarial regularizer that introduces a sensitive information removal function in the learning objective. The method is evaluated on three different primary tasks (identity, attractiveness, and smiling) and three publicly available benchmarks. In addition, we present a new face annotation dataset with balanced distribution between genders and ethnic origins. The experiments demonstrate that it is possible to improve the privacy and equality of opportunity while retaining competitive performance independently of the task.

The world is becoming more interconnected every day. With the high technological evolution and the increasing deployment of it in our society, scenarios based on the Internet of Things (IoT) can be considered a reality nowadays. However, and before some predictions become true (75 billion devices are expected to be interconnected in the next few years), many efforts must be carried out in terms of scalability and security. In this study we propose and evaluate a new approach based on the incorporation of Blockchain into current IoT scenarios. The main contributions of this study are as follows: i) an in-depth analysis of the different possibilities for the integration of Blockchain into IoT scenarios, focusing on the limited processing capabilities and storage space of most IoT devices, and the economic cost and performance of current Blockchain technologies; ii) a new method based on a novel module named BIoT Gateway that allows both unidirectional and bidirectional communications with IoT devices on real scenarios, allowing to exchange any kind of data; and iii) the proposed method has been fully implemented and validated on two different real-life IoT scenarios, extracting very interesting findings in terms of economic cost and execution time. The source code of our implementation is publicly available in the Ethereum testnet.

In this study we estimate the heart rate from face videos for student assessment. This information could be very valuable to track their status along time and also to estimate other data such as their attention level or the presence of stress that may be caused by cheating attempts. The recent edBBplat, a platform for student behavior modelling in remote education, is considered in this study1. This platform permits to capture several signals from a set of sensors that capture biometric and behavioral data: RGB and near infrared cameras, microphone, EEG band, mouse, smartwatch, and keyboard, among others. In the experimental framework of this study, we focus on the RGB and near-infrared video sequences for performing heart rate estimation applying remote photoplethysmography techniques. The experiments include behavioral and physiological data from 25 different students completing a collection of tasks related to e-learning. Our proposed face heart rate estimation approach is compared with the heart rate provided by the smartwatch, achieving very promising results for its future deployment in e-learning applications.

This work proposes and analyzes the use of keystroke biometrics for content de-anonymization. Fake news have become a powerful tool to manipulate public opinion, especially during major events. In particular, the massive spread of fake news during the COVID-19 pandemic has forced governments and companies to fight against missinformation. In this context, the ability to link multiple accounts or profiles that spread such malicious content on the Internet while hiding in anonymity would enable proactive identification and blacklisting. Behavioral biometrics can be powerful tools in this fight. In this work, we have analyzed how the latest advances in keystroke biometric recognition can help to link behavioral typing patterns in experiments involving 100,000 users and more than 1 million typed sequences. Our proposed system is based on Recurrent Neural Networks adapted to the context of content de-anonymization. Assuming the challenge to link the typed content of a target user in a pool of candidate profiles, our results show that keystroke recognition can be used to reduce the list of candidate profiles by more than 90%. In addition, when keystroke is combined with auxiliary data (such as location), our system achieves a Rank-1 identification performance equal to 52.6% and 10.9% for a background candidate list composed of 1K and 100K profiles, respectively.

The increased need for unattended authentication in multiple scenarios has motivated a wide deployment of biometric systems in the last few years. This has in turn led to the disclosure of security concerns specifically related to biometric systems. Among them, presentation attacks (PAs, i.e., attempts to log into the system with a fake biometric characteristic or presentation attack instrument) pose a severe threat to the security of the system: any person could eventually fabricate or order a gummy finger or face mask to impersonate someone else. In this context, we present a novel fingerprint presentation attack detection (PAD) scheme based on i) a new capture device able to acquire images within the short wave infrared (SWIR) spectrum, and ii) an in-depth analysis of several state-of-theart techniques based on both handcrafted and deep learning features. The approach is evaluated on a database comprising over 4700 samples, stemming from 562 different subjects and 35 different presentation attack instrument (PAI) species. The results show the soundness of the proposed approach with a detection equal error rate (D-EER) as low as 1.35% even in a realistic scenario where five different PAI species are considered only for testing purposes (i.e., unknown attacks).

Passwords are still used on a daily basis for all kind of applications. However, they are not secure enough by themselves in many cases. This work enhances password scenarios through two-factor authentication asking the users to draw each character of the password instead of typing them as usual. The main contributions of this study are as follows: i) We present the novel MobileTouchDB public database, acquired in an unsupervised mobile scenario with no restrictions in terms of position, posture, and devices. This database contains more than 64K on-line character samples performed by 217 users, with 94 different smartphone models, and up to 6 acquisition sessions. ii) We perform a complete analysis of the proposed approach considering both traditional authentication systems such as Dynamic Time Warping (DTW) and novel approaches based on Recurrent Neural Networks (RNNs). In addition, we present a novel approach named Time-Aligned Recurrent Neural Networks (TA-RNNs). This approach combines the potential of DTW and RNNs to train more robust systems against attacks. A complete analysis of the proposed approach is carried out using both MobileTouchDB and e-BioDigitDB databases. Our proposed TA-RNN system outperforms the state of the art, achieving a final 2.38% Equal Error Rate, using just a 4-digit password and one training sample per character. These results encourage the deployment of our proposed approach in comparison with traditional typed-based password systems where the attack would have 100% success rate under the same impostor scenario.

This work enhances traditional authentication systems based on Personal Identification Numbers (PIN) and One-Time Passwords (OTP) through the incorporation of biometric information as a second level of user authentication. In our proposed approach, users draw each digit of the password on the touchscreen of the device instead of typing them as usual. A complete analysis of our proposed biometric system is carried out regarding the discriminative power of each handwritten digit and the robustness when increasing the length of the password and the number of enrolment samples. The new e-BioDigit database, which comprises on-line handwritten digits from 0 to 9, has been acquired using the finger as input on a mobile device. This database is used in the experiments reported in this work and it is available together with benchmark results in GitHub.11.https://github.com/BiDAlab/eBioDigitDB. Finally, we discuss specific details for the deployment of our proposed approach on current PIN and OTP systems, achieving results with Equal Error Rates (EERs) ca. 4.0 percent when the attacker knows the password. These results encourage the deployment of our proposed approach in comparison to traditional PIN and OTP systems where the attack would have 100 percent success rate under the same impostor scenario.

Data have become one of the most valuable things in this new era where deep learning technology seems to overcome traditional approaches. However, in some tasks, such as the verification of handwritten signatures, the amount of publicly available data is scarce, what makes difficult to test the real limits of deep learning. In addition to the lack of public data, it is not easy to evaluate the improvements of novel approaches compared with the state of the art as different experimental protocols and conditions are usually considered for different signature databases. To tackle all these mentioned problems, the main contribution of this study is twofold: i) we present and describe the new DeepSignDB on-line handwritten signature biometric public database, and ii) we propose a standard experimental protocol and benchmark to be used for the research community in order to perform a fair comparison of novel approaches with the state of the art. The DeepSignDB database is obtained through the combination of some of the most popular on-line signature databases, and a novel dataset not presented yet. It comprises more than 70K signatures acquired using both stylus and finger inputs from a total of 1526 users. Two acquisition scenarios are considered, office and mobile, with a total of 8 different devices. Additionally, different types of impostors and number of acquisition sessions are considered along the database. The DeepSignDB and benchmark results are available in GitHub.

This paper proposes a novel approach for on-line signature complexity detection based on Recurrent Neural Networks (RNNs). Complexity of handwritten signatures can vary from very simple ones (just a simple flourish) to very complex signatures (including the handwritten full name and complex flourish). Three different complexity levels are proposed: low, medium, and high. Time functions are extracted from the on-line signatures and a system based on RNNs (BLSTM in particular) is trained to classify the three levels of complexity over a ground truth manually labelled database (BiosecurID with 400 subjects). This initial model is used to automatically label a very large database (DeepSignDB) containing over 1500 subjects, which is then used to train the proposed RNN for signature complexity detection. Promising results ca. 85% of accuracy are achieved. This complexity detector could be used as a first stage in a signature verification system in order to train a specific biometric system per signature complexity level and improve the overall system performance.

BioTouchPass enhances traditional authentication systems based on Personal Identification Numbers (PIN) and One-Time Passwords (OTP) through the incorporation of biometric information from handwriting as a second level of user authentication. In our proposed approach, users draw each digit of the password on the touchscreen of the device instead of typing them as usual. This way the security of the authentication system increases as impostors need more than the traditional password to get access to the system. BioTouchPass achieves results with Equal Error Rates (EERs) ca. 4.0% when the attacker knows the password, outperforming other authentication schemes based on touch biometrics, and providing a user-friendly interface easily adaptable to a variety of mobile devices and application scenarios.

Architectures based on Recurrent Neural Networks (RNNs) have been successfully applied to many different tasks such as speech or handwriting recognition with state-of-the art results. The main contribution of this work is to analyse the feasibility of RNNs for on-line signature verification in real practical scenarios. We have considered a system based on Long Short-Term Memory (LSTM) with a Siamese architecture whose goal is to learn a similarity metric from pairs of signatures. For the experimental work, the BiosecurID database comprised of 400 users and 4 separated acquisition sessions are considered. Our proposed LSTM RNN system has outperformed the results of recent published works on the BiosecurID benchmark in figures ranging from 17.76% to 28.00% relative verification performance improvement for skilled forgeries.

On-line signature verification systems are mainly based on two approaches: feature- or time functions-based systems (a.k.a. global and local systems). However, new sources of information can be also considered in order to complement these traditional approaches, reduce the intra-class variability and achieve more robust signature verification systems against forgers. In this paper we focus on the use of the concept of complexity in on-line signature verification systems. The main contributions of the present work are: 1) classification of users according to the complexity level of their signatures using features extracted from the Sigma LogNormal writing generation model, and 2) a new architecture for signature verification exploiting signature complexity that results in highly improved performance. Our proposed approach is tested considering the BiosecurID on-line signature database with a total of 400 users. Results of 5.8% FRR for a FAR = 5.0% have been achieved against skilled forgeries outperforming recent related works. In addition, an analysis of the optimal time functions for each complexity level is performed providing practical insights for the application of signature verification in real scenarios.