Bernadette Dorizzi

Hybrid template update system for unimodal biometric systems

Semi-supervised template update systems allow to automatically take into account the intra-class variability of the biometric data over time. Such systems can be inefficient by including too many impostors samples or skipping too many genuines samples. In the first case, the biometric reference drifts from the real biometric data and attracts more often impostors. In the second case, the biometric reference does not evolve quickly enough and also progressively drifts from the real biometric data. We propose a hybrid system using several biometric sub-references in order to increase performance of self-update systems by reducing the previously cited errors. The proposition is validated for a keystroke-dynamics authentication system (this modality suffers of high variability over time) on two consequent datasets from the state of the art.

A review on the public benchmark databases for static keystroke dynamics

Keystroke dynamics allows to authenticate individuals through their way of typing their password or a free text on a keyboard. In general, in biometrics, a novel algorithm is validated through a comparison to the state of the art ones using some datasets in an offline way. Several benchmark datasets for keystroke dynamics have been proposed in the literature. They differ in many ways and their intrinsic properties influence the performance of the algorithms under evaluation. In this work, we (a) provide a literature review on existing benchmark datasets of keystroke dynamics; (b) present several criteria and tests in order to characterize them; (c) and apply these criteria on these available public benchmark datasets. The review analysis shows a great disparity in the acquisition protocol, the population involved, the complexity of the passwords, or the expected performance (there is a relative difference of 76% between the EER on the worst and best performing datasets with the same authentication method).

Effective elliptic fitting for iris normalization

Highlights? Increase in recognition performance achieved with an effective contour fitting scheme. ? Elliptic contour fitting based on Active Contours formulation is described. ? Proposed contour fitting method is not dependent of the segmentation algorithm. ? Results are confirmed on several public databases. Having an accurate parametric description of the iris borders is a critical issue for iris recognition systems based on Daugmans rubber sheet normalization. Many methods in the literature use very powerful and effective schemes for iris segmentation but often apply a simple estimator procedure, such as the Hough Transform or Least Square Fitting to get this parametric description. Those fitting methods are very sensitive to the segmentation quality as inaccuracies will provoke large errors in the resulting contour.In this article we propose an effective way to find optimal parameters for ellipses in order to proceed the normalization. Our method is based on a variational formulation of the well-known Active Contour techniques leading to a compact formulation for elliptic contours. We show improvements compared to an Elliptic Hough Transform and a Direct Least Square Fitting on the following databases: ICE2005, ND-Iris and Casia-Lamp. We also demonstrate that our scheme can be paired effectively with different segmentation algorithms. Significant improvements of the recognition results were obtained when adding our algorithm after the segmentation stage of VASIR and OSIRIS, two open source packages for iris recognition.

Predicting postoperative gait in cerebral palsy

In this work, postoperative lower limb kinematics are predicted with respect to preoperative kinematics, physical examination and surgery data. Data of 115 children with cerebral palsy that have undergone single-event multilevel surgery were considered. Preoperative data dimension was reduced utilizing principal component analysis. Then, multiple linear regressions with 80% confidence intervals were performed between postoperative kinematics and bilateral preoperative kinematics, 36 physical examination variables and combinations of 9 different surgical procedures. The mean prediction errors on test vary from 4° (pelvic obliquity and hip adduction) to 10° (hip rotation and foot progression), depending on the kinematic angle. The unilateral mean sizes of the confidence intervals vary from 5° to 15°. Frontal plane angles are predicted with the lowest errors, however the same performance is achieved when considering the postoperative average signals. Sagittal plane angles are better predicted than transverse plane angles, with statistical differences with respect to the average postoperative kinematics for both planes angles except for ankle dorsiflexion. The mean prediction errors are smaller than the variability of gait parameters in cerebral palsy. The performance of the system is independent of the preoperative state severity of the patient. Even if the system is not yet accurate enough to define a surgery plan, it shows an unbiased estimation of the most likely outcome, which can be useful for both the clinician and the patient. More patients data are necessary for improving the precision of the model in order to predict the kinematic outcome of a large number of possible surgeries and gait patterns.

Watermarking signal fusion in multimodal biometrics

In this paper, we propose a new approach based on watermarking for fusing biometric modalities. The main idea of the proposed approach is to use the watermark both for security purpose and as an additional information related to the person, therefore increasing the personal data used for verification. The host image of the face is watermarked in the multi-resolution space by the palmprint using a watermarking technique based on wavelet packet decomposition. For the verification stage, the characterization of the watermarked face is provided by Gabor filters while classification is performed by SVMs. The experimental results show that this technique ensures a significant performance improvement in both identity verification and biometric security over the use of a single system.

Impact of Quality-Based Fusion Techniques for Video-Based Iris Recognition at a Distance

In this paper, we consider the problem of iris recognition in the context of video-based distant acquisition. We propose several systems aiming at improving the poor performance resulting from image degradations (low resolution, blur, and lack of texture) obtained from such acquisitions. Our approach is based on simple super-resolution techniques applied at the pixel level on the different frames of a video, improved by considering some quality criteria. Our main novelty is the introduction of a local quality measure in the fusion scheme. This measure relies on a gaussian mixture model estimation of clean iris texture distribution. It can also be used to compute a global quality measure of the normalized iris image which can be used either for the selection of the best images in a sequence or in the fusion scheme. Extensive experiments on the QFIRE database at different acquisition distances (5, 7, and 11 ft) show the big improvement brought by the use of the global quality for both scenarios. Moreover, the local quality-based fusion scheme further increases the performance due to its ability to consider locally the different parts of the image, and therefore, to discard poorly segmented pixels in the fusion.

C-FMCW Based Contactless Respiration Detection Using Acoustic Signal

Recent advances in ubiquitous sensing technologies have exploited various approaches to monitoring vital signs. One of the vital signs is human respiration which typically requires reliable monitoring with low error rate in practice. Previous works in respiration monitoring however either incur high cost or suffer from poor error rate. In this paper, we propose a Correlation based Frequency Modulated Continuous Wave method (C-FMCW) which is able to achieve high ranging resolution. Based on C-FMCW, we present the design and implementation of an audio-based highly-accurate system for human respiration monitoring, leveraging on commodity speaker and microphone widely available in home environments. The basic idea behind the audio-based method is that when a user is close to a pair of speaker and microphone, body movement during respiration causes periodic audio signal changes, which can be extracted to obtain the respiration rate. However, several technical challenges exist when applying C-FMCW to detect respiration with commodity acoustic devices. First, the sampling frequency offset between speakers and microphones if not being corrected properly would cause high ranging errors. Second, the uncertain starting time difference between the speaker and microphone varies over time. Moreover, due to multipath effect, weak periodic components due to respiration can easily be overwhelmed by strong static components in practice. To address those challenges, we 1) propose an algorithm to compensate dynamically acoustic signal and counteract the offset between speaker and microphone; 2) co-locate speaker and microphone and use the received signal without reflection (self-interference) as a reference to eliminate the starting time difference; and 3) leverage the periodicity of respiration to extract weak periodic components with autocorrelation. Extensive experimental results show that our system detects respiration in real environments with the median error lower than 0.35 breaths/min, outperforming the state-of-the-arts.

Quality-Based Super Resolution for Degraded Iris Recognition

In this paper we address the problem of low-quality iris recognition via super resolution approaches. We introduce two novel quality measures, onecomputed Globally (GQ) and the other Locally (LQ), for fusing at the pixel level(after a bilinear interpolation step) the images corresponding to several shots of a given person. These measures derive from a local GMM probabilistic characterization of good quality iris texture. We performed two types of experiments. The first oneconsiders low resolution video sequences coming from the MBGC portal database: it shows the superiority of our approach compared to score-based or average image-based fusion methods. Moreover, we show that the LQ-based fusionoutperforms the GQ-based fusion with a relative improvement of 4.79 % at the Equal Error Rate functioning point. The second experiment is performed on CASIA v4database containing sequences of still images with degraded quality resulting insevere segmentation errors. We show that the image fusion scheme improves greatly the performance and that the LQ-based fusion is mainly interesting for low FAR values.

Two Bioinspired Methods for Dynamic Signatures Analysis

This work focuses on the problem of dynamic signature segmentation and representation. A brief review of segmentation techniques for online signatures and movement modelling is provided. Two dynamic signature segmentation/representation methods are proposed. These methods are based on psychophysical evidences that led to the well-known Minimum Jerk Model. These methods are alternatives to the existing techniques and are very simple to implement. Experimental evidence indicates that the Minimum Jerk is in fact a good choice for signature representation amongst the family of quadratic derivative cost functions defined in Section 2.

Obtaining Cryptographic Keys Using Multi-biometrics

Multi-biometric systems have several advantages over uni-biometrics based systems, such as, better verification accuracy, larger feature space to accommodate more subjects, and higher security against spoofing. Unfortunately, as in case of uni-biometric systems, multi-biometric systems also face the problems of nonrevocability, lack of template diversity, and possibility of privacy compromise. A combination of biometrics and cryptography is a good solution to eliminate these limitations. In this chapter we present a multi-biometric cryptosystem based on the fuzzy commitment scheme, in which, a crypto-biometric key is derived from multi-biometric data. An idea (recently proposed by the authors) denoted as FeaLingECc (Feature Level Fusion through Weighted Error Correction) is used for the multi-biometric fusion. The FeaLingECc allows fusion of different biometric modalities having different performances (e.g., face + iris). This scheme is adapted for a multi-unit system based on two-irises and a multi-modal system using a combination of iris and face. The difficulty in obtaining the crypto-biometric key locked in the system (and in turn the reference biometric data) is 189 bits for the two-iris system while 183 bits for the iris-face system using brute force attack. In addition to strong keys, these systems possess revocability and template diversity and protect user privacy.