http://dl.gi.de/handle/20.500.12116/37442 2026-07-23T11:25:43Z 2026-07-23T11:25:43Z Bisztray, Tamás Gruschka, Nils Bourlai, Thirimachos Fritsch, Lothar http://dl.gi.de/handle/20.500.12116/37475 2021-10-04T08:43:53Z 2021-01-01T00:00:00Z

Emerging biometric modalities and their use: Loopholes in the terminology of the GDPR and resulting privacy risks Bisztray, Tamás; Gruschka, Nils; Bourlai, Thirimachos; Fritsch, Lothar Brömme, Arslan; Busch, Christoph; Damer, Naser; Dantcheva, Antitza; Gomez-Barrero, Marta; Raja, Kiran; Rathgeb, Christian; Sequeira, Ana; Uhl, Andreas Technological advancements allow biometric applications to be more omnipresent than in any other time before. This paper argues that in the current EU data protection regulation, classification applications using biometric data receive less protection compared to biometric recognition. We analyse preconditions in the regulatory language and explore how this has the potential to be the source of unique privacy risks for processing operations classifying individuals based on soft traits like emotions. This can have high impact on personal freedoms and human rights and, therefore, should be subject to data protection impact assessment.

2021-01-01T00:00:00Z Dong, Xingbo Jin, Zhe Guo, Zhenhua Teoh, Andrew Beng Jin http://dl.gi.de/handle/20.500.12116/37474 2021-10-04T08:43:53Z 2021-01-01T00:00:00Z

Towards Generating High Definition Face Images from Deep Templates Dong, Xingbo; Jin, Zhe; Guo, Zhenhua; Teoh, Andrew Beng Jin Brömme, Arslan; Busch, Christoph; Damer, Naser; Dantcheva, Antitza; Gomez-Barrero, Marta; Raja, Kiran; Rathgeb, Christian; Sequeira, Ana; Uhl, Andreas Face recognition based on deep convolutional neural networks (CNN) has manifested superior accuracy. Despite the high discriminability of deep features generated by CNN, the vulnerability of the deep feature is often overlooked and leads to security and privacy concerns, particularly, the risks of reconstructing face images from the deep templates. In this paper, we propose a method to generate high definition (HD) face images from deep features. To be specific, the deep features extracted from CNN are mapped to the input (latent vector) of the pre-trained StyleGAN2 using a regression model. Subsequently, HD face images can be generated based on the latent vector by the pre-trained StyleGAN2 model. To evaluate our method, we derived the face features from the generated HD face images and compared against the bona fide face features. In the sense of face image reconstruction, our method is simple, yet the experimental results suggest the effectiveness, which achieves an attack performance as high as TAR=46.08% (18.30%) @ FAR=0.1 threshold under type-I (type-II) attack settings. Besides, experiment results also indicate that 50.7% of generated HD face images can pass one commercial off-the-shelf (COTS) liveness detection.

2021-01-01T00:00:00Z Kantarcı, Alperen Dertli, Hasan Ekenel, Hazım Kemal http://dl.gi.de/handle/20.500.12116/37473 2021-10-04T08:43:52Z 2021-01-01T00:00:00Z

Shuffled Patch-Wise Supervision for Presentation Attack Detection Kantarcı, Alperen; Dertli, Hasan; Ekenel, Hazım Kemal Brömme, Arslan; Busch, Christoph; Damer, Naser; Dantcheva, Antitza; Gomez-Barrero, Marta; Raja, Kiran; Rathgeb, Christian; Sequeira, Ana; Uhl, Andreas Face anti-spoofing is essential to prevent false facial verification by using a photo, video, mask, or a different substitute for an authorized person's face. Most of the state-of-the-art presentation attack detection (PAD) systems suffer from overfitting, where they achieve near-perfect scores on a single dataset but fail on a different dataset with more realistic data. This problem drives researchers to develop models that perform well under real-world conditions. This is an especially challenging problem for frame-based presentation attack detection systems that use convolutional neural networks (CNN). To this end, we propose a new PAD approach, which combines pixel-wise binary supervision with patch-based CNN. We believe that training a CNN with face patches allows the model to distinguish spoofs without learning background or dataset-specific traces. We tested the proposed method both on the standard benchmark datasets ---Replay-Mobile, OULU-NPU--- and on a real-world dataset. The proposed approach shows its superiority on challenging experimental setups. Namely, it achieves higher performance on OULU-NPU protocol 3, 4 and on inter-dataset real-world experiments.

2021-01-01T00:00:00Z Rúa, Enrique Argones van Hamme, Tim Preuveneers, Davy Joosen, Wouter http://dl.gi.de/handle/20.500.12116/37472 2021-10-04T08:43:52Z 2021-01-01T00:00:00Z

Gait Authentication based on Spiking Neural Networks Rúa, Enrique Argones; van Hamme, Tim; Preuveneers, Davy; Joosen, Wouter Brömme, Arslan; Busch, Christoph; Damer, Naser; Dantcheva, Antitza; Gomez-Barrero, Marta; Raja, Kiran; Rathgeb, Christian; Sequeira, Ana; Uhl, Andreas In this paper we address gait authentication using a novel approach based on spiking neural networks (SNNs). This technology has proven advantages regarding energy consumption and it is a perfect match with some proposed neuromorphic hardware chips, which can lead to a broader adoption of user device applications of artificial intelligence technologies. One of the challenges when using this technology is the training of the network itself, since it is not straightforward to apply well-known error backpropagation, massively used in traditional artificial neural networks (ANNs). In this paper we propose a new derivation of error backpropagation for the spiking neural networks that integrates lateral inhibition and provides competitive results when compared to state of the art ANNs in the context of IMU-based gait authentication.

2021-01-01T00:00:00Z