Accepted papers are

Suppressing Gender and Age in Face Templates Using Incremental Variable Elimination
Authors: Philipp Terhörst, Naser Damer, Florian Kirchbuchner, Arjan Kuijper (all Fraunhofer IGD)
Abstract: Recent research on soft-biometrics showed that more information than just the persons identity can be deduced from biometric data. Using face templates only, information about gender, age, ethnicity, sexual orientation and the health state of the person can be automatically obtained. Since for most applications these template are expected to be used for recognition purposes only, this raises major privacy issues. Previous work addressed this problem purely on image level regarding function creep attackers without knowledge about the systems privacy mechanism. In this work, we propose a soft-biometric privacy enhancing approach that reduces a given biometric template by eliminating its most important variables for predicting softbiometric attributes. Training a decision tree ensemble allows to derive a variable importance measure that is used to incrementally eliminate variables that allow to predict sensitive attributes. Unlike previous work, we consider a scenario of function creep attackers with explicit knowledge about the privacy mechanism and evaluated our approach on a public available database. The experiments were conducted to eight baseline solutions. The results showed that in many cases IVE is able to suppress gender and age to a high degree with a negligible loss of the templates recognition ability. Contrary to previous work, which is limited to the suppression of binary (gender) attributes, IVE is able to suppress binary, categorical, and continuous attributes.

To Detect or not to Detect: The Right Faces to Morph
Authors: Naser Damer, Alexandra Moseguí Saladié, Steffen Zienert, Yaza Wainakh, Philipp Terhörst, Florian Kirchbuchner, Arjan Kuijper (all Fraunhofer IGD)
Abstract: Recent works have studied the face morphing attack detection performance generalization over variations in morphing approaches, image re-digitization, and image source variations. However, these works assumed a constant approach for selecting the images to be morphed (pairing) across their training and testing data. A realistic variation in the pairing protocol in the training data can result in problems and opportunities for a stable attack detector. This work extensively study this issue by building a novel database with three different pairing protocols and two different morphing approaches. We study the detection generalization over these variations for single image and differential attack detection, along with handcrafted and CNN-based features. Our observations included that training an attack detection solution on attacks created from dissimilar face images, in contrary to the common practice, can result in an overall more generalized detection performance. Moreover, we found that differential attack detection is very sensitive to variations in morphing and pairing protocols.

Thermal and Cross-spectral Palm Image Matching in the Visual Domain by Robust Image Transformation
Authors: Ewelina Bartuzi (NASK, Poland), Naser Damer (Fraunhofer IGD)
Abstract: Synthesizing visual-like images from those captured in the thermal spectrum allows for direct cross-domain comparisons. Moreover, it enables thermal-to-thermal comparisons that take advantage of feature extraction methodologies developed for the visual domain. Hand based biometrics are socially accepted and can operate in a touchless mode. However, certain deployment scenarios requires captures in non-visual spectrums due to impractical illumination requirements. Generating visual-like palm images from thermal ones faces challenges related to the nature of hand biometrics. Such challenges are the dynamic nature of the hand and the related difficulties in accurately aligning hand's scale and rotation, especially in the understudied thermal domain. Building such a synthetic solution is also challenged by the lack of large-scale databases that contain images collectec in both spectra, as well as producing images of appropriate resolutions. Driven by these challenges, this paper presents a novel solution to transfer thermal palm images into high-quality visual-like images, regardless of the limited training data, or scale and rotational variations. We proved quality similarity and high correlation of the generated images to the original visual images. We used the synthesized images within verification approaches based on CNN and hand crafted-features. This allowed significantly improved the cross-spectral and thermal-to-thermal verification performances, reducing the EER from 37.12% to 16.25% and from 3.04% to 1.65%, respectively in both cases when using CNN-based features.

Cross-spectrum thermal to visible face recognition based on cascaded image synthesis
Authors: Khawla Mallat (EURECOM, France), Naser Damer (Fraunhofer IGD), Fadi Boutros (Fraunhofer IGD), Arjan Kuijper (Fraunhofer IGD), Jean-Luc Dugelay (EURECOM, France)
Abstract: Face synthesis from thermal to visible spectrum is fundamental to perform cross-spectrum face recognition as it simplifies its integration in existing commercial face recognition systems and enables manual face verification. In this paper, a new solution based on cascaded refinement networks is proposed. This method generates visible-like colored images of high visual quality without requiring large amounts of training data. By employing a contextual loss function during training, the proposed network is inherently scale and rotation invariant. We discuss the visual perception of the generated visible-like faces in comparison with recent works. We also provide an objective evaluation in terms of cross-spectrum face recognition, where the generated faces were compared against a gallery in visible spectrum using two state-of-the-art deep learning based face recognition systems. When compared to the recently published TV-GAN solution, the performance of the face recognition systems, OpenFace and LightCNN, was improved by a 42.48% and a 71.43%, respectively.

On the Impact of Different Fabrication Materials on Fingerprint Presentation Attack Detection
Authors: Lazaro J. Gonzalez-Soler (CETANAV, Cuba), Marta Gomez-Barrero (HDA), Leonardo Chang (TEC, Mexico), Airel Perez Suarez CETANAV, Cuba), Christoph Busch (HDA)
Abstract: Presentation Attack Detection (PAD) is the task of determining whether a sample stems from a live subject (bona fide presentation) or from an artificial replica (Presentation Attack Instrument, PAI). Several PAD approaches have shown high effectiveness to successfully detect PAIs when the materials used for the fabrication of these PAIs are known a priori. However, most of these PAD methods do not take into account the characteristics of PAIs’ species in order to generalise to new, realistic and more challenging scenarios, where materials might be unknown. Based on that fact, in this work, we explore the impact of different PAI species, fabricated with different materials, on several local-based descriptors combined with the Fisher Vector feature encoding, in order to increase the robustness to unknown attacks. The experimental results over the well- established benchmarks of the LivDet 2011, LivDet 2013 and LivDet 2015 competitions reported error rates outperforming the top state-of-the-art in the presence of unknown attacks. Moreover, the evaluation revealed the differences in the detection performance due to the variability between the PAI species.

Multi-Modal Fingerprint Presentation Attack Detection: Analysing the Surface and the Inside
Autthors: Marta Gomez-Barrero (HDA), Jascha Kolberg (HDA), Christoph Busch (HDA)
Abstract: The deployment of biometric recognition systems has seen a considerable increase over the last decade, in particular for fingerprint based systems. To tackle the security issues derived from presentation attacks launched on the biometric capture device, automatic presentation attack detection (PAD) methods have been proposed. In spite of their high detection rates on the LivDet databases, the vast majority of the methods rely on the samples provided by traditional capture devices, which may fail to detect more sophisticated presentation attack instrument (PAI) species. In this paper, we propose a multi-modal fingerprint PAD which relies on an analysis of: i) the surface of the finger within the short wave infrared (SWIR) spectrum, and ii) the inside of the finger thanks to the laser speckle contrast imaging (LSCI) technology. On the experimental evaluation over a database comprising more than 4700 samples and 35 PAI species, and including unknown attacks to model a realistic scenario, a Detection Equal Error Rate (D-EER) of 0.5% has been achieved. More- over, for a BPCER ≤ 0.1% (i.e., highly convenient system), the APCER remains around 3%.

ICB 2019 is taking place in Kreta, Greece, from July 4th - 7th, 2019.

Information about ICB 2019

show all news