A5059766088- Face recognition and analysis
- Biometric Identification and Security
- Face and Expression Recognition
- User Authentication and Security Systems
- Facial Nerve Paralysis Treatment and Research
- Generative Adversarial Networks and Image Synthesis
- Face Recognition and Perception
- Digital Media Forensic Detection
- Anomaly Detection Techniques and Applications
- Forensic Fingerprint Detection Methods
- Video Surveillance and Tracking Methods
- Domain Adaptation and Few-Shot Learning
- Forensic and Genetic Research
- Advanced Image and Video Retrieval Techniques
- Privacy-Preserving Technologies in Data
- Context-Aware Activity Recognition Systems
- COVID-19 diagnosis using AI
- Network Security and Intrusion Detection
- Facial Rejuvenation and Surgery Techniques
- Gait Recognition and Analysis
- Visual Attention and Saliency Detection
- Retinal Imaging and Analysis
- Vehicle License Plate Recognition
- Digital and Cyber Forensics
- Gaze Tracking and Assistive Technology
Fraunhofer Institute for Computer Graphics Research
2016-2025
Technical University of Darmstadt
2019-2024
Versus Arthritis
2022
Fraunhofer Society
2013-2021
IAP Research (United States)
2018
Idiap Research Institute
2018
University of Kaiserslautern
2011
Systems incorporating biometric technologies have become ubiquitous in personal, commercial, and governmental identity management applications. Both cooperative (e.g. access control) non-cooperative surveillance forensics) systems benefited from biometrics. Such rely on the uniqueness of certain biological or behavioural characteristics human beings, which enable for individuals to be reliably recognised using automated algorithms. Recently, however, there has been a wave public academic...
Learning discriminative face features plays a major role in building high-performing recognition models. The recent state-of-the-art solutions proposed to incorporate fixed penalty margin on commonly used classification loss function, softmax loss, the normalized hypersphere increase power of models, by minimizing intra-class variation and maximizing inter-class variation. Marginal losses, such as ArcFace CosFace, assume that geodesic distance between within different identities can be...
Face image quality is an important factor to enable high-performance face recognition systems. assessment aims at estimating the suitability of a for purpose recognition. Previous work proposed supervised solutions that require artificially or human labelled values. However, both labelling mechanisms are error prone as they do not rely on clear definition and may know best characteristics utilized system. Avoiding use inaccurate labels, we novel concept measure based arbitrary model. By...
Biometric recognition technology has made significant advances over the last decade and is now used across a number of services applications. However, this widespread deployment also resulted in privacy concerns evolving societal expectations about appropriate use technology. For example, ability to automatically extract age, gender, race, health cues from biometric data heightened leakage. Face technology, particular, been spotlight, seen by many as posing considerable risk personal...
Face image quality assessment (FIQA) estimates the utility of captured in achieving reliable and accurate recognition performance. This work proposes a novel FIQA method, CR-FIQA, that face sample by learning to predict its relative classifiability. classifiability is measured based on allocation training feature representation angular space with respect class center nearest negative center. We experimentally illustrate correlation between As such property only observable for dataset, we...
This article presents FRCSyn-onGoing, an ongoing challenge for face recognition where researchers can easily benchmark their systems against the state of art in open common platform using large-scale public databases and standard experimental protocols. FRCSyn-onGoing is based on Face Recognition Challenge Era Synthetic Data (FRCSyn) organized at WACV 2024. first international aiming to explore use real synthetic data independently, also fusion, order address existing limitations technology....
Face morphing attacks aim at creating face images that are verifiable to be the of multiple identities, which can lead building faulty identity links in operations like border crossing. Research has been focused on more accurate attack detection approaches by considering different image properties. However, all considered so far based manipulating facial landmarks localized morphed images. In contrast, this work presents novel generated generative adversarial networks. We present MorGAN...
Sensors are devices that quantify the physical aspects of world around us. This ability is important to gain knowledge about human activities. Human Activity recognition plays an import role in people's everyday life. In order solve many human-centered problems, such as health care, and individual assistance, need infer various simple complex activities prominent. Therefore, having a well defined categorization sensing technology essential for systematic design activity systems. By extending...
Face recognition (FR) systems have a growing effect on critical decision-making processes. Recent works shown that FR solutions show strong performance differences based the user's demographics. However, to enable trustworthy technology, it is essential know influence of an extended range facial attributes beyond Therefore, in this work, we analyze bias over wide attributes. We investigate 47 verification two popular models. The experiments were performed publicly available MAAD-Face...
The primary objective of face morphing is to com-bine images different data subjects (e.g. an malicious actor and accomplice) generate a image that can be equally verified for both contributing subjects. In this paper, we propose new framework generating morphs using newer Generative Adversarial Network (GAN) - StyleGAN. contrast earlier works, realistic high-quality high resolution 1024 × pixels. With the newly created dataset 2500 morphed images, pose critical question in work. (i) Can GAN...
Face morphing attacks target to circumvent Recognition Systems (FRS) by employing face images derived from multiple data subjects (e.g., accomplices and malicious actors). Morphed can be verified against contributing with a reasonable success rate, given they have high degree of facial resemblance. The is directly dependent on the quality generated morph images. We present new approach for generating strong extending our earlier framework morphs. using an Identity Prior Driven Generative...
Since early 2020, the COVID-19 pandemic has had a considerable impact on many aspects of daily life. A range different measures have been implemented worldwide to reduce rate new infections and manage pressure national health services. primary strategy gatherings potential for transmission through prioritisation remote working education. Enhanced hand hygiene use facial masks decreased spread pathogens when are unavoidable. These particular present challenges reliable biometric recognition,...
Deep neural networks have rapidly become the mainstream method for face recognition (FR). However, this limits deployment of such models that contain an extremely large number parameters to embedded and low-end devices. In work, we present lightweight accurate FR solution, namely PocketNet. We utilize architecture search develop a new family face-specific architectures. additionally propose novel training paradigm based on knowledge distillation (KD), multi-step KD, where is distilled from...
Recent deep face recognition models proposed in the literature utilized large-scale public datasets such as MS-Celeb-1M and VGGFace2 for training very neural networks, achieving state-of-the-art performance on mainstream benchmarks. Recently, many of these datasets, e.g., VGGFace2, are retracted due to credible privacy ethical concerns. This motivates this work propose investigate feasibility using a privacy-friendly synthetically generated dataset train models. Towards end, we utilize...
The main question this work aims at answering is: "can morphing attack detection (MAD) solutions be successfully developed based on synthetic data?". Towards that, introduces the first synthetic-based MAD development dataset, namely Synthetic Morphing Attack Detection Development dataset (SMDD). This is utilized to train three backbones where it proved lead high performance, even completely unknown types. Additionally, an essential aspect of detailed legal analyses challenges using and...
In this work, we propose QMagFace, a simple and effective face recognition solution (QMagFace) that combines quality-aware comparison score with model based on magnitude-aware angular margin loss. The proposed approach includes model-specific image qualities in the process to enhance performance under unconstrained circumstances. Exploiting linearity between their scores induced by utilized loss, our function is highly generalizable. experiments conducted several databases benchmarks...
Over the past years, main research innovations in face recognition focused on training deep neural networks large-scale identity-labeled datasets using variations of multi-class classification losses. However, many these are retreated by their creators due to increased privacy and ethical concerns. Very recently, privacy-friendly synthetic data has been proposed as an alternative privacy-sensitive authentic comply with regulations ensure continuity research. In this paper, we propose...
The availability of large-scale authentic face databases has been crucial to the significant advances made in recognition research over past decade. However, legal and ethical concerns led recent retraction many these by their creators, raising questions about continuity future without one its key resources. Synthetic datasets have emerged as a promising alternative privacy-sensitive data for development. synthetic that are used train models suffer either from limitations intra-class...
Despite the widespread adoption of face recognition technology around world, and its remarkable performance on current benchmarks, there are still several challenges that must be covered in more detail. This paper offers an overview Face Recognition Challenge Era Synthetic Data (FRCSyn) organized at WACV 2024. is first international challenge aiming to explore use synthetic data address existing limitations technology. Specifically, FRCSyn targets concerns related privacy issues, demographic...
As a crucial security problem, anti-spoofing in biometrics, and particularly for the face modality, has achieved great progress recent years. Still, new threats arrive inform of better, more realistic sophisticated spoofing attacks. The objective 2nd Competition on Counter Measures to 2D Face Spoofing Attacks is challenge researchers create counter measures effectively detecting variety submitted propositions are evaluated Replay-Attack database results presented this paper.
In this paper, we present a set of extremely efficient and high throughput models for accurate face verification, Mix-FaceNets which are inspired by Mixed Depthwise Convolutional Kernels. Extensive experiment evaluations on Label Face in the Wild (LFW), Age-DB, MegaFace, IARPA Janus Benchmarks IJB-B IJB-C datasets have shown effectiveness our MixFaceNets applications requiring low computational complexity. Under same level computation complexity (≤ 500M FLOPs), outperform MobileFaceNets all...