P196 - BIOSIG 2012 - Proceedings of the 11th International Conference of the Biometrics Special Interest Group
Auflistung nach:
Auflistung P196 - BIOSIG 2012 - Proceedings of the 11th International Conference of the Biometrics Special Interest Group nach Erscheinungsdatum
1 - 10 von 37
Treffer pro Seite
Sortieroptionen
- KonferenzbeitragImproving channel robustness in text-independent speaker verification using adaptive virtual cohort models(BIOSIG 2012, 2012) Nautsch, Andreas; Schönwandt, Anne; Kasper, Klaus; Reininger, Herbert; Wagner, MartinIn speaker verification, score normalization methods are a common practice to gain better performance and robustness. One kind of score normalization is cohort normalization, which uses information about the score behaviour of known impostors. During enrolment, impostor verifications are simulated to get a speaker-specific set of the most competitive impostors (the cohort). In the present paper, one virtual cohort speaker is synthesized using the most competitive impostor's Hidden Markov Models (HMMs). These impostors are also users of the system and therefore their models have channel-specific information contrary to the universal background model, which provides channeland speaker-independent models. On verification, cohort scores are obtained by an additional verification of the virtual cohort speaker. The cohort scores evaluate the candidate as an impostor. A cohort normalized score promises greater robustness. This paper will study the effect of the introduced cohort normalization technique on the speaker verification system atip VoxGuard, which is based on mel-frequency cepstral coefficients and HMMs. VoxGuard can be used as either a text-dependent or a text-independent verification system. In this paper, emphasis is placed on textindependent speaker verification. Experiments using the atip speech corpus and the SieTill speech corpus showed improvements measured by the equal error rate on performance and robustness.
- KonferenzbeitragCommon mistakes on face recognition based on video(BIOSIG 2012, 2012) Despiegel, Vincent; Gentric, StéphaneAutomatic border control e gates relies on face recognition in video streams. Two family of methods could be imagined to define a common workflow: encode all that is possible or select the best image of the stream according to a quality assessment. In gates, giving an answer as soon as possible by encoding each and every possible image seems to lead to faster gates. Is it really the case?
- KonferenzbeitragImage metric-based biometric comparators: a supplement to feature vector-based Hamming distance?(BIOSIG 2012, 2012) Hofbauer, Heinz; Rathgeb, Christian; Uhl, Andreas; Wild, PeterIn accordance with the ISO/IEC FDIS 19794-6 standard an iris-biometric fusion of image metric-based and Hamming distance (HD) comparison scores is presented. In order to demonstrate the applicability of a knowledge transfer from image quality assessment to iris recognition, Peak Signal to Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), Local Edge Gradients metric (LEG), Edge Similarity Score (ESS), Local Feature Based Visual Security (LFBVS), and Visual Information Fidelity (VIF) are applied to iris textures, i.e. query textures are interpreted as noisy representations of registered ones. Obtained scores are fused with traditional HD scores obtained from iris-codes generated by different feature extraction algorithms. Experimental evaluations on the CASIA-v3 iris database confirm the soundness of the proposed approach.
- KonferenzbeitragEvaluation of automatic face recognition for automatic border control on actual data recorded of travellers at Schiphol Airport(BIOSIG 2012, 2012) Spreeuwers, Luuk J.; Hendrikse, Anne J.; Gerritsen, Kier-Co J.Automatic border control at airports using automated facial recognition for checking the passport is becoming more and more common. A problem is that it is not clear how reliable these automatic gates are. Very few independent studies exist that assess the reliability of automated facial recognition for border control. In this evaluation study the reliability of automated facial recognition for automatic border passage was investigated. To investigate the quality of the images and face recognition, during 2 weeks data of real passengers were acquired at Schiphol Airport using 2 different automatic gates of about 950 passengers for both gates. This data alone already makes the evaluation study of great value. The evaluation experiment consisted of comparing live images of every passenger to the digital photographs stored on their passports. Every live image is compared to every digital passport photograph. In this way we can estimate both the False Accept Rate as well as the Verification Rate. In spite of the critical analysis in this study, the prospects for automatic border passage using face recognition are very good. We expect that, provided that the quality of the live images acquired by the gates is improved and if possible the quality of the digital photographs stored on the passport, excellent recognition results can be obtained with Verification Rates (VR) of above 99% at a False Accept Rate (FAR) of 0.1% or even lower.
- KonferenzbeitragProbabilistic matching pair selection for SURF-based person re-identification(BIOSIG 2012, 2012) Khedher, Mohamed Ibn; El-Yacoubi, Mounim A.; Dorizzi, BernadetteThe objective of this paper is to study the performance of human reidentification based on multi-shot SURF and to assess its degradation according to the angular difference between the test and reference video scene view angles. In this context, we propose a new automatic statistical method of acceptance and rejection of SURF correspondence based on the likelihood ratio of two GMMs learned on the reference set and modeling the distribution of distances resulting from matching sequences associated with the same person and with different persons respectively. The experimental results show that our approach compares favorably with the state of the art and achieves a good performance.
- KonferenzbeitragChallenge-based speaker recognition for mobile authentication(BIOSIG 2012, 2012) Baloul, Mossah; Cherrier, Estelle; Rosenberger, ChristopheUser authentication is a major trend to guarantee the security of electronic transactions when using mobile devices such as tablets or mobile phones. Biometrics is for us the only real user authentication method. In this article, we propose to realize a speaker recognition approach to achieve this goal. We use a challenge-based method to avoid the replay attack (especially if the impostor has recorded the user's voice). In this case, free text recognition is realized. Experimental results on the CMU database show very good results, while providing low computation times.
- KonferenzbeitragUnsupervised learning of face detection models from unlabeled image streams(BIOSIG 2012, 2012) Walther, Thomas; Würtz, Rolf P.Modern artificial face detection shows impressive performance in a variety of application areas. This success comes at the cost of supervised training, using large-scale databases provided by human experts. In this paper, we propose a face detection system based on Organic Computing [vdM08] paradigms that acquires necessary domain knowledge autonomously and learns a conceptual model of the human face/head region. Performance of the novel approach is experimentally compared to state-of-the-art face detection, yielding competitive results in scenarios of moderate complexity.
- KonferenzbeitragEvaluation of security for biometric guessing attacks in biometric cryptosystem using fuzzy commitment scheme(BIOSIG 2012, 2012) Hidano, Seira; Ohki, Tetsushi; Takahashi, KentaBiometric authentication based on template protection has attracted attention in the past decade. In the discussion on the security of these systems, however, the content of biometric information is assumed to be sufficiently large, and real conditions of biometric features have not yet been reflected. This paper focuses on a biometric cryptosystem using a fuzzy commitment scheme and demonstrates correlation between fingerprint bit strings by using our method for evaluating the content of biometric information. Additionally, attacks to guess biometric bit strings, which take advantage of correlation between them, are explained, and the security against these attacks is theoretically and experimentally discussed.
- KonferenzbeitragBiometrics for an ageing society – societal and ethical factors in biometrics and ageing(BIOSIG 2012, 2012) Rebera, Andrew P.; Guihen, BarryBy the middle of the twenty-first century around one third of the European population will be aged 65 or over. This poses two main challenges to biometrics. First, the quality of an image capturable from an older person is likely to be inferior to that of a younger person, leading to increased failure to capture or failure to enroll rates. Second, since biometric features alter over time, `within-person variation' and `template ageing' lead to significant system performance degradation. As society ages the need for solutions becomes increasingly urgent. This paper addresses a major societal and ethical issue this need provokes.
- KonferenzbeitragA topology based approach to categorization of fingerprint images(BIOSIG 2012, 2012) Aabrandt, Andreas; Olsen, Martin Aastrup; Busch, ChristophThis paper discusses the use of betti numbers to characterize fingerprint and iris images. The goal is to automatically separate fingerprint images from nonfingerprint images; where non-fingerprint images of special interest are biometric samples which are not fingerprints. In this regard, an image is viewed as a triangulated point cloud and the topology associated with this construct is summarized using its first betti number - a number that indicates the number of distinct cycles in the triangulation associated to the particular image. This number is then compared against the first betti numbers of “n” prototype images in order to perform classification (“fingerprint” vs “non-fingerprint”). The proposed method is compared against SIVV (a tool provided by NIST). Experimental results on fingerprint and iris databases demonstrate the potential of the scheme.