Advancement of machine learning techniques, combined with the availability of large-scale datasets, has significantly improved the accuracy and efficiency of facial recognition. Modern facial recognition systems are trained using large face datasets collected from diverse individuals or public repositories. However, for training, these datasets are often replicated and stored in multiple workstations, resulting in data replication, which complicates database management and oversight. Currently, once a user submits their face for dataset preparation, they lose control over how their data is used, raising significant privacy and ethical concerns. This paper introduces VOIDFace, a novel framework for facial recognition systems that addresses two major issues. First, it eliminates the need of data replication and improves data control to securely store training face data by using visual secret sharing. Second, it proposes a patch-based multi-training network that uses this novel training data storage mechanism to develop a robust, privacy-preserving facial recognition system. By integrating these advancements, VOIDFace aims to improve the privacy, security, and efficiency of facial recognition training, while ensuring greater control over sensitive personal face data. VOIDFace also enables users to exercise their Right-To-Be-Forgotten property to control their personal data. Experimental evaluations on the VGGFace2 dataset show that VOIDFace provides Right-To-Be-Forgotten, improved data control, security, and privacy while maintaining competitive facial recognition performance.

This work summarises and reports the results of the second Presentation Attack Detection competition on ID cards. This new version includes new elements compared to the previous one. (1) An automatic evaluation platform was enabled for automatic benchmarking; (2) Two tracks were proposed in order to evaluate algorithms and datasets respectively; and (3) A new ID card dataset was shared with Track 1 teams to serve as the baseline dataset for the training and optimisation. The Hochschule Darmstadt, Fraunhofer-IGD, and Facephi company jointly organised this challenge. 20 teams were registered, and 74 submitted models were evaluated. For Track 1, the "Dragons" team reached first place with an Average Ranking and Equal Error rate (EER) of (AVRank) of 40.48% and 11.44% EER, respectively. For the more challenging approach in Track 2, the "Incode" team reached the best results with an AVRank of 14.76% and 6.36% EER, improving on the results of the first edition of 74.30% and 21.87% EER, respectively. These results suggest that PAD on ID cards is improving, but it is still a challenging problem related to the number of images, especially of bona fide images.

Adversarial attacks pose a significant threat to the reliability of biometric systems, particularly in security-critical applications such as identity verification and access control. Ensuring robustness against such attacks is essential for the safe deployment of face recognition technologies in real-world scenarios. To advance this goal, the 2025 Adversarial Attack Challenge for Secure Face Recognition was organized as part of the International Joint Conference on Biometrics (IJCB) 2025.The competition focused on two main tracks: Detection, where the objective was to determine whether a given face image is clean or adversarial, and Resilience, which aimed to evaluate recognition systems under adversarial perturbations. Participants were provided with a standardized dataset derived from CelebA and LFW, encompassing both clean samples and adversarial images crafted using ten diverse attack methods targeting evasion and impersonation scenarios. To ensure fairness and reproducibility, all models were trained solely on the data provided, with support from a custom open source adversarial attack package tailored for face recognition.In addition to benchmarking adversarial robustness, the challenge contributes to the research community by releasing the data set and the extensible attack package, allowing further investigation of secure and reliable face recognition systems.

Morphing is a long-standing problem in vision and computer graphics, requir- ing a time-dependent warping for feature alignment and a blending for smooth interpolation. Recently, multilayer perceptrons (MLPs) have been explored as implicit neural representations (INRs) for modeling such deformations, due to their meshlessness and differentiability; however, extracting coherent and accurate morphings from standard MLPs typically relies on costly regularizations, which often lead to unstable training and prevent effective feature alignment. To overcome these limitations, we propose FLOWING (FLOW morphING), a framework that recasts warping as the construction of a differential vector flow, naturally ensuring continuity, invertibility, and temporal coherence by encoding structural flow prop- erties directly into the network architectures. This flow-centric approach yields principled and stable transformations, enabling accurate and structure-preserving morphing of both 2D images and 3D shapes. Extensive experiments across a range of applications—including face and image morphing, as well as Gaussian Splatting morphing—show that FLOWING achieves state-of-the-art morphing quality with faster convergence. Code and pretrained models are available in https://schardong.github.io/flowing.

Covariance matrices outperform first-order features in many tasks, attracting considerable attention from the computer vision research community. Covariance matrices encode second-order statistics between features, at the same time it is robust to noise. Based on this, we propose representing images by covariance matrices and defining a loss function that measures the distance between them through the Riemannian distance. Motivated by the robustness and invariance properties of the affine invariant Riemannian metric the proposed method was validated in printer-proof data transmission, which is a challenging task due to the trade-off between image quality and message recovery capabilities after printing and digitization procedures. The effectiveness of this approach was systematically assessed using MS COCO and IMM Face datasets. The results demon-strated that the proposed approach outperforms conventional methods that use Euclidean distance, generating encoded images with better quality and achieving higher recovery accuracy in printed images. Additionally, a broader application of the proposed loss was successfully tested in image generation tasks, using generative adversarial networks (GANs).

Face morphing is a problem in computer graphics with numerous artistic and forensic applications. It is challenging due to variations in pose, lighting, gender, and ethnicity. This task consists of a warping for feature alignment and a blending for a seamless transition between the warped images. We propose to leverage coord-based neural networks to represent such warpings and blendings of face images. During training, we exploit the smoothness and flexibility of such networks by combining energy functionals employed in classical approaches without discretizations. Additionally, our method is time-dependent, allowing a continuous warping/blending of the images. During morphing inference, we need both direct and inverse transformations of the time-dependent warping. The first (second) is responsible for warping the target (source) image into the source (target) image. Our neural warping stores those maps in a single network dismissing the need for inverting them. The results of our experiments indicate that our method is competitive with both classical and generative models under the lens of image quality and face-morphing detectors. Aesthetically, the resulting images present a seamless blending of diverse faces not yet usual in the literature.

Face recognition has achieved outstanding performance in the last decade with the development of deep learning techniques. Nowadays, the challenges in face recognition are related to specific scenarios, for instance, the performance under diverse image quality, the robustness for aging and edge cases of person age (children and elders), distinguishing of related identities. In this set of problems, recognizing children's faces is one of the most sensitive and important. One of the reasons for this problem is the existing bias towards adults in existing face datasets. In this work, we present a benchmark dataset for children's face recognition, which is compiled similarly to the famous face recognition benchmarks LFW, CALFW, CPLFW, XQLFW and AgeDB. We also present a development dataset (separated into train and test parts) for adapting face recognition models for face images of children. The proposed data is balanced for African, Asian, Caucasian, and Indian races. To the best of our knowledge, this is the first standardized data tool set for benchmarking and the largest collection for development for children's face recognition. Several face recognition experiments are presented to demonstrate the performance of the proposed data tool set.

Imagine a world where a copy of your face could trick the most advanced security systems. This isn’t science fiction; it’s a real challenge today. LivDet-Face is a competition that aims to advance the detection of attacks at the biometric sensor, known as Presentation Attack Detection (PAD). This international contest is a key benchmark in biometric security, offering an unbiased look at the latest innovations in face PAD and demonstrating progress over time in detecting and preventing sophisticated attacks. Through the International Joint Conference on Biometrics (IJCB) platform, LivDet-Face 2024 provides a standardized evaluation process, access to advanced Presentation Attack Instruments (PAI), and a comprehensive dataset of bona fide face images. The competition had two main categories: algorithms and systems. A total of sixteen algorithms and one system were submitted for this year’s competition. Anonymous submissions topped both image and video subcategories with an ACER of 4.93% and 4.13%, respectively. In the systems category, Team Dermalog, despite being the sole submission, achieved an impressive ACER of 3.12%.

Biometrics in the realm of face image modality, has seen significant advancements in recent decades, which was driven by the rise of deep learning techniques. With widespread deployment across various domains, including document security and user authentication, face recognition based systems are increasingly susceptible to presentation attacks. In this work we address the issue of estimating the robustness of face recognition systems to face morphing attacks. We revisit the definition of Mated Morph Presentation Match Rate metrics and develop the benchmarking utilities for these metrics on the novel dataset. Through extensive experiments conducted with our benchmark, we estimate the robustness of various public face recognition models to face morphing attacks. Furthermore, we evaluate the efficiency of different face morphing techniques in deceiving face recognition systems.

Facial phenotypes are extensively studied in medical and biological research, serving as critical markers that potentially indicate underlying genetic traits or medical conditions. With the recent advancements in big data, algorithms, and hardware, deep facial diagnosis, which employs deep learning techniques to systematically examine facial phenotypes and identify signs of certain diseases or medical conditions, has attracted significant attention and research, gradually emerging as a promising tool in precision medicine. Primarily limited by the scarcity of data for training facial diagnosis models, the accuracy of facial diagnosis for various conditions remains low up to now. In the past decade, RGB-D cameras, measuring depth information along with standard RGB capabilities, have proven superior in processing spatial details with more stability and accuracy. Motivated by the facts mentioned above, in this paper, we propose a Simulated Multimodal Framework, which effectively improves the computer-aided facial diagnosis performance of state-of-the-art models in experiments under different conditions. The underlying principle is to leverage the simulated depth by generative models to improve the performance of RGB image recognition. Furthermore, as a rapid and non-invasive tool for disease screening and detection, our proposal demonstrated an accuracy improvement of over 20% compared to practicing physicians in the study.

Advancements in deep learning techniques and availability of large scale face datasets led to significant performance gains in face recognition in recent years. Modern face recognition algorithms are trained on large-scale in-the-wild face datasets. At the same time, many facial biometric applications rely on controlled image acquisition and enrollment procedures (for instance, document security applications). That is why such face recognition approaches can demonstrate the deficiency of the performance in the target scenario (ICAO-compliant images). However, modern approaches for face image quality estimation may help to mitigate that problem. In this work, we introduce a strategy for filtering training datasets by quality metrics and demonstrate that it can lead to performance improvements in biometric applications that rely on face image modality. We filter the main academic datasets using the proposed filtering strategy and present performance metrics.

Dissertação de Mestrado Integrado em Engenharia Electrotécnica e de Computadores apresentada à Faculdade de Ciências e Tecnologia

Face morphing attack detection (MAD) is one of the most challenging tasks in the field of face recognition nowadays. In this work, we introduce a novel deep learning strategy for a single image face morphing detection, which implies the discrimination of morphed face images along with a sophisticated face recognition task in a complex classification scheme. It is directed onto learning the deep facial features, which carry information about the authenticity of these features. Our work also introduces several additional contributions: the public and easy-to-use face morphing detection benchmark and the results of our wild datasets filtering strategy. Our method, which we call MorDeephy, achieved the state of the art performance and demonstrated a prominent ability for generalizing the task of morphing detection to unseen scenarios.

Object detection in autonomous driving applications implies the detection and tracking of semantic objects that are commonly native to urban driving environments, as pedestrians and vehicles. One of the major challenges in state-of-the-art deep-learning based object detection are false positives which occur with overconfident scores. This is highly undesirable in autonomous driving and other critical robotic-perception domains because of safety concerns. This paper proposes an approach to alleviate the problem of overconfident predictions by introducing a novel probabilistic layer to deep object detection networks in testing. The suggested approach avoids the traditional Sigmoid or Softmax prediction layer which often produces overconfident predictions. It is demonstrated that the proposed technique reduces overconfidence in the false positives without degrading the performance on the true positives. The approach is validated on the 2D-KITTI objection detection through the YOLOV4 and SECOND (Lidar-based detector). The proposed approach enables interpretable probabilistic predictions without the requirement of re-training the network and therefore is very practical.

Depth map, also known as range image, can directly reflect the geometric shape of the objects. Due to several issues such as cost, privacy and accessibility, face depth information is not easy to obtain. However, the spatial information of faces is very important in many aspects of computer vision especially in the biometric identification. In contrast, scene depth information is related easier to obtain with the development of autonomous driving technology in recent years. An idea of face depth estimation inspired is to bridge the gap between the scene depth and the face depth. Previously, face depth estimation and scene depth estimation were treated as two completely separate domains. This paper proposes and explores utilizing scene depth knowledge learned to estimate the depth map of faces from monocular 2D images. Through experiments, we have preliminarily verified the possibility of using scene depth knowledge to predict the depth of faces and its potential in face feature representation.

Light-field cameras are becoming more popular in the consumer market. Their data redundancy allows, in theory, to accurately refocus images after acquisition and to predict the depth of each point visible from the camera. Combined, these two features allow for the generation of full-focus images, which is impossible in traditional cameras. Multiple methods for depth prediction from light fields (or stereo) have been proposed over the years. A large subset of these methods relies on cost-volume estimates - 3D objects where each layer represents a heuristic of whether each point in the image is at a certain distance from the camera. Generally, this volume is used to regress a depth map, which is then refined for better results. In this paper, we argue that refining the cost volumes is superior to refining the depth maps in order to further increase the accuracy of depth predictions. We propose a set of cost-volume refinement algorithms and show their effectiveness.

Light field cameras capture a scene's multi-directional light field with one image, allowing the estimation of depth. In this paper, we introduce a fully automatic method for depth estimation from a single plenoptic image running a RANSAC-like algorithm for feature matching. The novelty about our method is the global method to back project correspondences found using photometric similarity to obtain a 3D virtual point cloud and different methods to build a depth map from the 3D point cloud generated. We use lenses with different focal-lengths in a multiple depth map refining phase, generating a dense depth map. Tests with simulations and real images are presented and compared with the state of the art, showing comparable accuracy for substantial less computational time.

In this paper we present a framework for the application of augmented reality to a mobile robot, using non-central camera systems. Considering a virtual object in the world with known local 3D coordinates, the goal is to project this object into the image of a non-central catadioptric imaging device. We propose a solution to this problem which allows us to project textured objects to the image in real-time (up to 20 fps): projection of 3D segments to the image; occlusions; and illumination. In addition, since we are considering that the imaging device is on a mobile robot, one needs to take into account the real-time localization of the robot. To the best of our knowledge this is the first time that this problem is addressed (all state-of-the-art methods are derived for central camera systems). To evaluate the proposed framework we test the solution using a mobile robot and a non-central catadioptric camera (using a spherical mirror).

In this paper, we address the problem of pose estimation under the framework of generalized camera models. We propose a solution based on the knowledge of the coordinates of 3-D straight lines (expressed in the world coordinate frame) and their corresponding image pixels. Previous approaches used the knowledge of the coordinates of 3-D points (zero dimensional elements) and their corresponding images (zero dimensional elements). In this paper, pixels belonging to the image of 3-D lines are used. There is no need to establish correspondences between pixels and 3-D points. Correspondences are established between 3-D lines and their images. There is no need to identify individual pixels. The use of correspondences between pixels (that belong to the images of the 3-D lines) and 3-D lines facilitates the correspondence problem when compared to the use of world and image points. This is one of the contributions of this paper. The approach is both evaluated and validated using synthetic data and also real images.

The tradeoff between error-correcting performance and numerical complexity of LDPC decoding algorithms is a well-known problem. In this paper we depict the unseen error-floor performance of the Self-Corrected Min-Sum algorithm for long length DVB-S2 codes. We developed a massively parallel simulation using GPUs which allowed a comprehensive BER characterization either in the waterfall or in the error-floor region. We show that the self-correction technique increases the BER performance by 0.5 and 0.2 dB, in the waterfall and error-floor region, when compared to the Min-Sum algorithm. Furthermore, it reaches within 0.2 dB to the Logarithmic Sum-Product BER performance and it also outperforms the Normalized Min-Sum at high SNR, a low complexity decoding algorithm which yields good BER performance.

Estimating the amount and center of distortion from lines in the scene has been addressed in the literature by the so-called ``plumb-line'' approach. In this paper we propose a new geometric method to estimate not only the distortion parameters but the entire camera calibration (up to an ``angular'' scale factor) using a minimum of 3 lines. We propose a new framework for the unsupervised simultaneous detection of natural image of lines and camera parameters estimation, enabling a robust calibration from a single image. Comparative experiments with existing automatic approaches for the distortion estimation and with ground truth data are presented.

Recovering the epipolar geometry of a stereo image pair is important for many computer and robotic vision systems, for performing motion recovering, 3D reconstruction and, more recently, image retrieval from large databases. Most state-of-the-art methods for estimating the fundamental matrix rely solely in putative image correspondences, and, therefore, heavily depend on the capability of the low-level image features to provide enough distinctiveness capabilities for establishing correct matches. In this paper we present a robust method for estimating the fundamental matrix based on all image features, and not only matching points. This is done by selecting the best correspondent keypoints between views through a proper weighting function that fuses local appearance of keypoints and distance to the epipolar lines. Several distance weighting functions are compared, with an intuitive theoretical analysis of the role of each function parametrization being analyzed. Experimental evidence shows that our approach outperforms the current state-of-the-art methods in terms of error magnitude, number of correct matches provided and computational time.

This paper presents a linear method to estimate the pose of a noncentral catadioptric system with a quadric shaped mirror in relation to a world reference frame (or local reference frame without loss of generality). The vision system is assumed to be calibrated. The method uses also as input data the structure of the scene. It is proved that any reflection point should belong to an analytical quadric that intersects the mirror quadric itself. This constraint can be written linearly in the 3D scene point coordinates (in the camera reference frame). The unknown pose screw transformation that relates camera and world reference frames can then be used in the linear model, allowing for the construction of a linear equation in the pose transformation elements. Additional constraints are used to force the estimated rotation elements to build an orthogonal matrix. Tests with simulated data and also on real images with different mirrors proved the method to be consistent and to estimate the pose accurately. However, it was also observed that the method is sensitive to noise. The results are compared with another method.

This paper analyses the problem of motion estimation from a sequence of stereo images. Both the differential and discrete approaches of two methods are formulated The differential approach uses differential optical flow whereas the discrete approaches uses feature correspondences. Both methods are used to compute, first, the 3D velocity in the depth (Z) direction and, second, the complete rigid motion parameters. Furthermore, the uncertainty propagation models for both methods and approaches are derived These models are analysed in order to point out the critical variables for the methods. The methods were extensively tested using synthetic images as well as real images and several conclusions are drawn from the results. The real images are used without any illumination control of the scene in order to study the behavior of the methods in strongly noisy environments with low resolution depth maps.