P230 - BIOSIG 2014 - Proceedings of the 13th International Conference of the Biometrics Special Interest Group
Auflistung P230 - BIOSIG 2014 - Proceedings of the 13th International Conference of the Biometrics Special Interest Group nach Erscheinungsdatum
1 - 10 von 26
Treffer pro Seite
Sortieroptionen
- KonferenzbeitragRe-assessing the threat of replay spoofing attacks against automatic speaker verification(BIOSIG 2014, 2014) Alegre, Federico; Janicki, Artur; Evans, NicholasThis paper re-examines the threat of spoofing or presentation attacks in the context of automatic speaker verification (ASV). While voice conversion and speech synthesis attacks present a serious threat, and have accordingly received a great deal of attention in the recent literature, they can only be implemented with a high level of technical know-how. In contrast, the implementation of replay attacks require no specific expertise nor any sophisticated equipment and thus they arguably present a greater risk. The comparative threat of each attack is re-examined in this paper against six different ASV systems including a state-of-the-art iVector-PLDA system. Despite the lack of attention in the literature, experiments show that low-effort replay attacks provoke higher levels of false acceptance than comparatively higher-effort spoofing attacks such as voice conversion and speech synthesis. Results therefore show the need to refocus research effort and to develop countermeasures against replay attacks in future work.
- KonferenzbeitragFixed FAR vote fusion of regional facial classifiers(BIOSIG 2014, 2014) Spreeuwers, Luuk J.; Veldhuis, Raymond N. J.; Sultanali, Siar; Diephuis, JasperHolistic face recognition methods like PCA and LDA have the disadvantage that they are very sensitive to expression, hair and illumination variations. This is one of the main reasons they are no longer competitive in the major benchmarks like FRGC and FRVT. In this paper we present an LDA based approach that combines many overlapping regional classifiers (experts) using what we call a Fixed FAR Voting Fusion (FFVF) strategy. The combination by voting of regional classifiers means that if there are sufficient regional classifiers unaffected by the expression, illumination or hair variations, the fused classifier will still correctly recognise the face. The FFVF approach has two interesting properties: it allows robust fusion of dependent classifiers and it only requires a single parameter to be tuned to obtain weights for fusion of different classifiers. We show the potential of the FFVF of regional classifiers using the standard benchmarks experiments 1 and 4 on FRGCv2 data. The multi-region FFVF classifier has a FRR of 4\% at FAR=0.1\% for controlled and 38\% for uncontrolled data compared to 7\% and 56\% for the best single region classifier.
- KonferenzbeitragSocial acceptance of epassports(BIOSIG 2014, 2014) Tiits, Marek; Kalvet, Tarmo; Laas-Mikko, KatrinUsing large-scale web survey in six countries we study the societal readiness and acceptance of specific technology options in relation to the potential next generation of ePassports. We find that the public has only limited knowledge of the electronic data and functions ePassports include, and often have no clear opinion on various potential uses for ePassports and related personal data. Still, the public expects from ePassports improvements in protection from document forgery, accuracy and reliability of the identification of persons, and protection from identity theft. The main risks the public associates with ePassports includes the possible use of personal information for purposes other than those initially stated, and covert surveillance. Compared to earlier studies, our research shows that issues of possible privacy invasion and abuse of information are much more perceived by the public. There is a weak correlation between a persons' level of knowledge about ePassports and their willingness to accept the use of advanced biometrics, such as fingerprints or eye iris images, in different identity management and identity checking scenarios. Furthermore, the public becomes more undecided about ePassport applications as we move from the basic state of the art towards more advanced biometric technologies in various scenarios. The successful pathway to greater acceptability of the use of advanced biometrics in ePassports should start from the introduction of perceivably high-benefit and lowrisk applications. As the public awareness is low, citizens' belief in government benevolence, i.e. the belief that the government acts in citizens' best interest, comes out as an important factor in the overall context.
- KonferenzbeitragAn unconstrained Activity Recognition Method using Smart Phones(BIOSIG 2014, 2014) Çelenli, Naciye; Seviş, Kamile N.; Esgin, Muhammed F.; Altundaǧ, Kemal; Uludaǧ, UmutIn this paper, we provide human activity recognition performance rates, using accelerometer and gyroscope signals acquired using smart phones. Covering seven basic actions (walking, running, jumping, standing, ascending stairs, descending stairs, and standing up and sitting down as one action) and a complex action (getting in and out of a car), with more than 100 subjects in a database collected in different environments, we provide recognition results on the largest database in the published literature. Utilizing features (e.g. extrema, zero crossing rates$\dots $) extracted from time-windows (e.g. with a duration of 2 seconds), K-Star classifier led to the best performance among 6 classifiers tested, exceeding 98\% recognition accuracy. A detailed comparison with current approaches is provided, along with possible future research directions. The associated technology could be helpful for health-related monitoring of one's activities, generating automatic status feeds for social networking sites, and calculating precise/adaptive calorie intake needs for individuals.
- KonferenzbeitragAssessing facial age similarity: A framework for evaluating the robustness of different feature sets(BIOSIG 2014, 2014) Lanitis, Andreas; Tsapatsoulis, NicolasA framework that can be used for assessing the suitability of different feature vectors in the task of determining the age similarity between a pair of faces is
- KonferenzbeitragShattering the glass maze(BIOSIG 2014, 2014) Hermans, Jens; Peeters, Roel; Mennink, BartTemplate protection plays a crucial role in protecting the privacy of biometric data, by providing irreversibility and unlinkability. The Glass Maze, as presented by Trugenberger at BIOSIG 2011, is a fingerprint key binding mechanism that is claimed to provide template protection. With the correct fingerprint, the key that is entangled with the fingerprint data can be retrieved. The template protection of the Glass Maze is based on the convergence properties of a Hopfield model, a neural network. We however show how to revert the Glass Maze to recover the key, without requiring a correct fingerprint. This completely breaks the irreversibility property, and hence also unlinkability.
- KonferenzbeitragEfficient two-stage speaker identification based on universal background models(BIOSIG 2014, 2014) Billeb, Stefan; Rathgeb, Christian; Buschbeck, Michael; Reininger, Herbert; Kasper, KlausConventional speaker identification systems are already field-proven with respect to recognition accuracy. Since any biometric identification requires exhaustive 1 : N comparisons for identifying a biometric probe, comparison time frequently dominates the overall computational workload, preventing the system from being executed in real-time. In this paper we propose a computational efficient two-stage speaker identification system based on Gaussian Mixture Model and Universal Background Model. Binarized voice biometric templates are utilized to pre-screen a large database and thereby reduce the required amount of full comparisons to a fraction of the total. Experimental evaluations demonstrate that the proposed system is capable of significantly accelerating the response-time of the system and, at the same time, identification performance is maintained, confirming the soundness of the scheme.
- KonferenzbeitragBimodal palm biometric feature extraction using a single RGB image(BIOSIG 2014, 2014) Eglitis, Teodors; Pudzs, Mihails; Greitans, ModrisThis paper proposes a method for palm bimodal biometric feature (vein and crease pattern) acquisition from a single RGB image. Typical bimodal biometric systems require combining infrared and visible images for this task. We use a single CMOS color sensor and a specific illumination comprising of two wavelengths to acquire the image. As a result each biometric modality is more pronounced in its own color channel. The image is processed by applying adapted matched filters with nonlinear modifications. Performance of the proposed method is evaluated against feature separation with optical band-pass and band-stop approach on a database of 64 people. The results show the average true positive rate is 70.6 \% for vein detection and 64.7 \% for crease detection, whereas in 14.8 \% and 9.29 \% of the cases feature of wrong modality is detected.
- KonferenzbeitragAutomatic speaker verification using nearest neighbor normalization (3N) on an ipad tablet(BIOSIG 2014, 2014) Khemiri, Houssemeddine; Usoltsev, Alexander; Legout, Marie-Christine; Petrovska-Delacrétaz, Dijana; Chollet, GérardThis paper describes the development, implementation and validation of an automatic speaker recognition system on an iPad tablet. A score normalization approach, referred as Nearest Neighbor Normalization (3N), is applied in order to improve the baseline speaker verification system. The system is evaluated on the MOBIO corpus and results show an absolute improvement of the HTER by more than 4\% when the score normalization is performed. A human-centered interface is implemented for the speaker recognition system and a survey is collected from 28 users in order to evaluate the application. The results showed that the users, familiar with touchscreen interface, found the application easy to learn and use.
- KonferenzbeitragPre-processing cascades and fusion in finger vein recognition(BIOSIG 2014, 2014) Kauba, Christoph; Reissig, Jakob; Uhl, AndreasPreprocessing and fusion techniques for finger vein recognition are investigated. An experimental study involving a set of preprocessing approaches shows the importance of selecting the appropriate single technique and the usefulness of cascading several different preprocessing methods for subsequent feature extraction of various types. Score level fusion is able to significantly improve recognition results, in particular when combining features describing complementary finger image properties.
- «
- 1 (current)
- 2
- 3
- »