Search

International audience; Segmentation of the iris is one of the key modules of an iris recognition system. For this reason, it is critical to predict failures of this module. In this article we propose a new set of segmentation quality metrics dedicated this problem. We assess the quality of our metrics based on their ability to predict the intrinsic recognition performance of a segmented image. A straightforward fusion procedure then allows generating a global segmentation quality score.

International audience; This paper concerns the implementation of our online signature verification system on a mobile device. Verification involves confirming or denying a person's claimed identity. Our system is based on a Hidden Markov Model and outputs two complementary scores: the first one is related to the likelihood given by the HMM of the claimed identity; the second one is related to the segmentation given by such an HMM on the input signature. A claimed identity is confirmed when the arithmetic mean of the two scores obtained on such an input signature is higher than a threshold. Also, a personal normalization of the local parameters of the signature is carried out to make the system robust to changes of platforms. A patent was submitted with special emphasis on the latter claim. This system is implemented on a mobile platform PDA Qtek 2020 ARM 400 MHz. An acquisition interface is developed allowing an enrollment step of a person by acquisition of 5 of his/her signatures, and a verification step of a given signature of a registered person. Enrolment speed depends on the complexity of the signature, while verification is performed in real time. Performance assessment of our system, carried out on two databases acquired on a PDA, shows a degradation of system performance on mobile platform compared to a fixed platform. In order to improve the performance in the case of mobility, we propose a strategy for enhancing the quality of the reference signatures at the enrolment phase.

International audience; We present a simple scheme with three main features: (1) it induces revocability in biometrics based user authentication systems, (2) protects the biometric information, and (3) improves the verification performance. The user's biometric feature vector is shuffled with a user specific shuffling key to transform into a revocable template. Comparison between different templates is carried out in the transformed domain. If the template is compromised, it can be replaced with another template obtained by changing the shuffling key. This scheme makes cross-matching between databases impossible by issuing different templates for different applications. The performance evaluation of the system is carried out on two modalities: iris and face using publicly available databases. This scheme significantly improves the verification performance of the underlying biometric system, e.g., it reduces the Equal Error Rate (EER) from 1.67% to 0.23% on the NIST-ICE iris database. The EER on the NIST-FRGCv2 face database reduces from 8.10% to zero

The analytic structure of the solution of an ordinary differential equation is intimately related to its integrability. The Painleve property, i.e., pure poles being the only movable singularities, allows the identification of new integrable dynamical systems. In this paper, we recall briefly the Ablowita-Ramani-Segur (ARS) algorithm

International audience; In this paper, we present some results on multimodality implementation resulting from the VINSI (“Vérification d’Identité Numérique Sécurisée Itinérante” for Secured Mobile Digital Identity Verification) French project. The VINSI handheld terminal allows identity verification in mobile conditions (airport gates) using two biometrics usable in future biometric passports (fingerprint and face). We propose an incremental fusion strategy aiming at improving the global performance of the combined system over each individual recognizer while optimizing the cost resulting from the fusion. Indeed, in this kind of application, time and complexity optimization is essential. To this aim, we split the fingerprint scores’ range into different interest zones, on which we do not apply the same strategy depending on the relative quality of the modalities at hand. Results on a virtual database corresponding to VINSI applicative conditions (Combination of BIOMET fingerprints and FRGCv2 faces) show that this incremental fusion strategy allows the same improvement in performance as global fusion methods while significantly reducing the cost

The purpose of this paper is to study the influence of face expressions on the performance of a 3D face recognition algorithm. Three facial surface matching based algorithms, namely ICP, Localized ICP (L-ICP) and Region-based ICP (R-ICP), are benchmarked on several sets of data : the two first sets with neutral faces and the last with expressive ones. Results show that the R-ICP algorithm provides more robustness to face expression verification than the two other approaches.

In this paper we propose a novel iris recognition method for iris images acquired under normal light illumination. We exploit the color information as we compare the distributions of common colors between a reference image and a test image using a modified Hausdorff distance. Tests have been made on the UBIRIS public database and on the IRIS_INT database acquired by our team. Comparisons with two iris reference systems in controlled scenario show a significant improvement when using color information instead of texture information. On uncontrolled scenarios, we propose a quality measure on colors in order to select good images from bad ones in the comparison process.

International audience; The home remote monitoring systems aim to make a protective contribution to the well being of individuals (patients, elderly persons) requiring moderate amounts of support for independent living spaces, and improving their everyday life. Existing researches of these systems suffer from lack of experimental data and a standard medical database intended for their validation and improvement. This paper presents a multi-sensors environment for acquiring and recording a multimodal medical database, which includes physiological data (cardiac frequency, activity or agitation, posture, fall), environment sounds and localization data. It provides graphical interface functions to manage, process and index these data. The paper focuses on the system implementation, its usage and it points out possibilities for future work

In this paper, we present a new method for iris recognition based on specific texture analysis. It relies on the use of Haar wavelet analysis to extract the texture of the iris tissue. The matching is done using a specific correlation based on local peaks detection. Experiments have been conducted on the CASIA database in verification mode and show an EER of 0.07%. Degraded version of the CASIA database results in an EER of 2.3%, which is lower than result obtained by the classical wavelet demodulation (WD) method in that database.

We propose in this paper a methodology for multibiometric systems evaluation on databases of virtual and real subjects of limited size (about 100 persons). Our study is limited to two biometric traits (modalities) that are a priori mutually independent, namely on-line signature and voice. Experiments are conducted on bimodal data of real subjects of the BIOMET database [9] and on several databases of virtual subjects constructed from BIOMET.

This paper describes an association of original hardware solutions associated to adequate software software for human face recognition. A differential CMOS imaging system [1] and a Synchronized flash camera [2] have been developed to provide ambient light invariant images and facilitate segmentation of the face from the background. This invariance of face image demonstrated by our prototype camera systems can result in a significant software/hardware simplification in such biometrics applications especially on a mobile platform where the computation power and memory capacity are both limited. In order to evaluate our prototypes we have build a face database of 25 persons with 4 different illumination conditions. These solutions with appropriate cameras give a significant improvement in performance (on the normal CCD cameras) using a simple correlation based algorithm associated with an adequate preprocessing. Finally, we have obtained a promising results using fusion between different sensors.

We describe a method fusing two complementary scores descended from a Hidden Markov Model (HMM) for on-line signature verification. The signatures are acquired using a digitizer that captures pen-position, pen-pressure, and pen-inclination. A writer is considered as being authentic when the arithmetic mean of two similarity scores obtained on an input signature is higher than a threshold. The first score is related to the likelihood given by a HMM modeling the signatures of the claimed identity; the second score is related to the most likely path given by such HMM (Viterbi algorithm) on the input signature. Our approach was evaluated on the BIOMET database (1266 genuine signatures from 87 individuals), as well as on the Philips on-line signature database (1530 signatures from 51 individuals). On the Philips database, we study the influence of the amount of training data, and on the BIOMET database, that of time variability. Several Cross-Validation trials are performed to report robust results. We first compare our system on the Philips database to Dolfing’s system, on one of his protocols (15 signatures to train the HMM). We reach in these conditions an Equal Error Rate (EER) of 0.95%, compared to an EER of 2.2% previously obtained by Dolfing. When considering only 5 signatures to train the HMM, the best results relying only on the likelihood yield an EER of 6.45% on the BIOMET database, and of 4.18% on the Philips database. The error rates drop to 2.84% on the BIOMET database, and to 3.54% on the Philips database, when fusing both scores by a simple arithmetic mean.

In this paper, we present a new method for iris identification particularly convenient for visible light images. It relies on the use of packets of wavelets [8] for the production of an code-vector. Experiments, conducted on a database of 700 iris images, acquired with visible light illumination, show an improvement of 1.5 % of FRR and of 12 % of FAR with the proposed method relatively to the classical wavelet method.

This work is concerned with the design and realization of a complete telecare application for remote monitoring of patients at home, including a wireless monitoring portable device held by the patient and a remote central server application located in a surveillance center. Agitation, position and cardiac rate data are used for alarm decision and can be exploited by the central server application. Originality resides in the actimetry and heart rate data combination for alarm decision and in the central server design aiming at offering flexible assistance services for both the surveillance medical team and the practitioner in charge of the patient.

Information technology innovations involve a constant evolution of man-machine interaction modes. Automated authentication of people could be used to better adapt the machine to the user. Security can also be enhanced through a better people authentication. Biometrics appears as a promising tool in these two situations. Different modalities can be envisaged, such as: fingerprint, human face images, hand shape, voice, handwritten signature... In order to take advantage of the particularities of each modality, and to improve the performance of a person authentication system, multimodality can be applied. This motivated the recording of BIOMET, a biometric database with five different modalities, including face, voice, fingerprint, hand and signature data. In this paper, the BIOMET multimodal database for person authentication is described. Details about the acquisition protocols of each modality are given. Preliminary monomodal verification results, obtained on a subcorpus of the BIOMET fingerprint data, are also presented.

We present a recurrent MLP1 neural network architecture as a predictive model for time-series caracterization and noise filtering. The approach has been tested on a periodic function corrupted by an additive white noise. Tests have been performed with several noise levels and the method is shown to outperform both linear ARMA2 models and other neural network approximators. Moreover, an improved implementation of the classical OBD3 pruning procedure allows an important reduction of the architecture of the network, without any loss in performance.

We present two connectionist modular approaches which are potentially able to deal with real applications as their size does not increase drastically with the size of the problem. The first model relics on a very simple cooperation of modular MLP networks specially designed for some sub-tasks. The second is based on a new methodology using a particular processing element (“neuron”) called yprel. The main characteristics of the approach are: (i) An yprel classifier is a set of yprel nets, each net being associated to a particular class; (ii) the learning is supervised and conducted class by class; (iii) the structure of the net is not a priori chosen, but is determined step by step during the learning process. Both approaches are compared on a well-known classification task (recognition of typographic characters) in terms of performance rates.