Mons, Belgium
Mons, Belgium
Time filter
Source Type

Larcher A.,Institute for Infocomm Research | Bonastre J.-F.,University of Avignon | Fauve B.,ValidSoft Ltd | Lee K.A.,Institute for Infocomm Research | And 4 more authors.
Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH | Year: 2013

ALIZE is an open-source platform for speaker recognition. The ALIZE library implements a low-level statistical engine based on the well-known Gaussian mixture modelling. The toolkit includes a set of high level tools dedicated to speaker recognition based on the latest developments in speaker recognition such as Joint Factor Analysis, Support Vector Machine, i-vector modelling and Probabilistic Linear Discriminant Analysis. Since 2005, the performance of ALIZE has been demonstrated in series of Speaker Recognition Evaluations (SREs) conducted by NIST and has been used by many participants in the last NISTSRE 2012. This paper presents the latest version of the corpus and performance on the NIST-SRE 2010 extended task. Copyright © 2013 ISCA.

Lecourt J.-B.,Multitel | Duterte C.,Multitel | Narbonneau F.,Multitel | Hernandez Y.,Multitel | Giannone D.,Multitel
Optics Express | Year: 2012

We report a PM all-normal, all-in-fiber passively mode-locked laser operating at 1030 nm. The main pulse shaping mechanism is provided by a tilted chirped-FBG. The laser delivers nanojoule range highly chirped pulses at a repetition rate of about 40 MHz. The FWHM of the optical spectrum is up to 7.8 nm leading to sub-500 fs compressed optical pulses. The influence of the filtering bandwidth and the output coupling ratio has been investigated. © 2012 Optical Society of America.

Boivinet S.,Multitel | Lecourt J.-B.,Multitel | Hernandez Y.,Multitel | Fotiadi A.A.,Electromagnetism and Telecommunications Unit | And 2 more authors.
IEEE Photonics Technology Letters | Year: 2014

We report an all-fiber passively mode-locked laser at 1030 nm. The principle of this ultrafast laser is based on nonlinear polarization evolution occurring in a long span of birefringent fiber spliced between a polarizer and a Faraday mirror. This configuration supports a large variety of pulsed regimes: noise-like, multiple pulses, burst, and single pulse mode-lock operation. In the single pulse mode-lock regime, the pulse duration is ∼ 17.8 ps at a low repetition rate of 948 kHz. The spectral bandwidth is ∼ 2.2 nm, the pulse energy is 68 pJ, and the polarization extinction ratio is estimated to 15 dB. © 2014 IEEE.

University of Limoges, Multitel and Leukos | Date: 2011-04-08

A supercontinuum source capable of emitting waves between the infrared range and the ultraviolet range, includes a pulsed laser source (12) capable of generating a laser beam and a non-linear microstructured optical fibre (14) capable of receiving the laser beam, in such a way that the supercontinuum source is capable of generating a beam (F) over a pulse duration (T), characterized in that the supercontinuum source includes elements for varying the pulse duration.

Witt J.,BAM Federal Institute of Materials Research and Testing | Narbonneau F.,Multitel | Schukar M.,BAM Federal Institute of Materials Research and Testing | Krebber K.,BAM Federal Institute of Materials Research and Testing | And 9 more authors.
IEEE Sensors Journal | Year: 2012

For patients under Magnetic Resonance Imaging (MRI) spontaneous respiration is constantly at risk of being impaired by anesthetic drugs or by upper airway obstruction. Therefore, continuous monitoring of the breathing activity is needed to assess adequate ventilation or to detect specific obstruction patterns. The paper describes three MRI compatible respiration sensors based on pure optical technologies developed within the EU FP6 project OFSETH. The sensors are based on fiber Bragg gratings, optical time-domain reflectometry and macrobending effects. The developed smart medical textiles can sense elongation up to 3% while maintaining the stretching properties of the textile substrates for patient's comfort. The OFSETH harness allows a continuous measurement of abdominal and thoracic respiration movement while all vitals organs are free for medical staff actions. The sensors were tested in MRI environment and on healthy adults. © 2006 IEEE.

Fernandez I.A.,Catholic University of Louvain | Chen F.,Catholic University of Louvain | Lavigney F.,University of Mons | Desurmontz X.,Multitel | De Vleeschouwer C.,Catholic University of Louvain
2nd International Conference on Advances in Multimedia, MMEDIA 2010 | Year: 2010

This paper builds on an interactive streaming architecture that supports both user feedback interpretation, and temporal juxtaposition of multiple video bitstreams in a single streaming session. As an original contribution, these functionalities can be exploited to offer improved viewing experience, when accessing football content through individual and potentially bandwidth constrained connections. Starting from a conventional broadcasted content, our system automatically splits the native content into non-overlapping and semantically consistent segments. Each segment is then divided into shots, based on conventional view boundary detection. Shots are finally splitted in small clips. These clips support our browsing capabilities during the whole playback in a temporally consistent way. Multiple versions are automatically created to render each clip. Versioning depends on the view type of the initial shot, and typically corresponds to the generation of zoomed in and spatially or temporally subsampled video streams. Clips are encoded independently so that the server can decide on the fly the version to send as a function of the semantic relevance of the segments (in a user-transparent basis, as inferred from video analysis or metadata) and the interactive user requests. Replaying certain game actions is also offered upon request. The streaming is automatically switched to the requested event. Later, the playback is resumed without any offset. The capabilities of our system rely on the H.264/AVC standard. We use soccer videos to validate our framework in subjective experiments showing the feasibility and relevance of our system. © 2010 IEEE.

Corbari C.,University of Southampton | Gladyshev A.V.,Fiber Optics Research Center | Lago L.,Multitel | Ibsen M.,University of Southampton | And 2 more authors.
Optics Letters | Year: 2014

All-fiber ns-pulsed visible laser at λ = 521 nm is realized by frequency doubling an Yb-doped fiber laser with a periodically poled silica fiber. A 50-mW second-harmonic (SH) output power is produced that is over 6-orders of magnitude greater than previous results obtained with poled fibers in the visible spectral range. The normalized conversion efficiency of 0.3 MW is to date the largest demonstrated with poled fiber technology. Furthermore, 21% conversion efficiency is achieved for the doubling of 8-ps pulses from a neodymium-doped yttrium vanadate solid-state laser. The advances are made possible by the precision and flexibility offered by using the continuous periodic UV erasure, as opposite to photolithographic methods, for the fabrication of over 20-cm-long χ2-gratings for quasi-phase matched SH generation. © 2014 Optical Society of America.

Boivinet S.,Multitel | Boivinet S.,University of Mons | Lecourt J.-B.,Multitel | Hernandez Y.,Multitel | And 2 more authors.
Proceedings of SPIE - The International Society for Optical Engineering | Year: 2014

We present in this study a PM all-fiber laser oscillator passively mode-locked (ML) at 1.03 μm. The laser is based on Nonlinear Polarization Evolution (NPE) in polarization maintaining (PM) fibers. In order to obtain the mode-locking regime, a nonlinear reflective mirror including a fibered polarizer, a long fiber span and a fibered Faraday mirror (FM) is inserted in a Fabry-Perot laser cavity. In this work we explain the principles of operation of this original laser design that permits to generate ultrashort pulses at low repetition (lower that 1MHz) rate with a cavity length of 100 m of fiber. In this experiment, the measured pulse duration is about 6 ps. To our knowledge this is the first all-PM mode-locked laser based on the NPE with a cavity of 100m length fiber and a delivered pulse duration of few picosecondes. Furthermore, the different mode-locked regimes of the laser, i.e. multi-pulse, noise-like mode-locked and single pulse, are presented together with the ways of controlling the apparition of these regimes. When the single pulse mode-locking regime is achieved, the laser delivers linearly polarized pulses in a very stable way. Finally, this study includes numerical results which are obtained with the resolution of the NonLinear Schrodinger Equations (NLSE) with the Split-Step Fourier (SSF) algorithm. This modeling has led to the understanding of the different modes of operation of the laser. In particular, the influence of the peak power on the reflection of the nonlinear mirror and its operation are studied. © 2014 SPIE.

Carincotte C.,Multitel | Bremond F.,French Institute for Research in Computer Science and Automation
Proceedings - IEEE International Conference on Advanced Video and Signal Based Surveillance, AVSS 2010 | Year: 2010

In Intelligent Video Systems, most of the recent advanced performance evaluation metrics perform a stage of mapping data between the system results and ground truth. This paper aims to review these metrics using a proposed framework. It will focus on metrics for events detection, objects detection and objects tracking systems. © 2010 IEEE.

Simon C.,Communication and Remote Sensing Laboratory | Meessen J.,Multitel | De Vleeschouwer C.,Communication and Remote Sensing Laboratory
Multimedia Tools and Applications | Year: 2010

This paper presents a classifier-based approach to recognize dynamic events in video surveillance sequences. The goal of this work is to propose a flexible event recognition system that can be used without relying on a long-term explicit tracking procedure. It is composed of three stages. The first one aims at defining and building a set of relevant features describing the shape and movements of the foreground objects in the scene. To this aim, we introduce new motion descriptors based on space-time volumes. Second, an unsupervised learning-based method is used to cluster the objects, thereby defining a set of coarse to fine local patterns of features, representing primitive events in the video sequences. Finally, events are modeled as a spatio-temporal organization of patterns based on an ensemble of randomized trees. In particular, we want this classifier to discover the temporal and causal correlations between the most discriminative patterns. Our system is experimented and validated both on simulated and real-life data. © 2009 Springer Science+Business Media, LLC.

Loading Multitel collaborators
Loading Multitel collaborators