The Information Technology research center organizes, as regularly as possible, scientific seminars. These seminar lunches are open to every researcher, from the research center or elsewhere.

Upcoming Seminars

By Sidi Mahmoudi (Computer Science Lab.), January 14th, 2010
GPU-based Image Processing with CUDA and OpenGL

Past Seminars

Slides of a seminar may be available upon electronic request to the author directly or to Alexis Moinet (firstname.lastname at umons.ac.be)

By Céline Mancas-Thillou (Transfer Technology Office), November 19th
Intellectual Property
By Thomas Drugman (TCTS Lab.) November 9th, 2009
Advances in glottal analysis and applications
Most of current speech processing systems rely on the information contained in the amplitude spectrum of the speech signal, related to the vocal tract contribution. Orthogonally to this approach, this talk will investigate the potential of using the glottal component of speech, which is generally neglected in these systems.
The presentation will be divided into two parts:
(a) the estimation of the glottal flow directly from speech audio recordings,
(b) the use of the glottal information in various applications such as speech synthesis, voice pathology detection, speaker recognition and expressive speech analysis.
By Serguei Iassinovski (Multitel), October 29th, 2009
Presentation of a smart RAO simulator
I will briefly introduce the concepts of the simulator, modelisation of complex discrete system, the RAO language, pedagogical models and some real-life study case of simulation/optimization. The simulator is based on an original idea of process presentation in terms of production rules (if - then) modified to consider the duration of actions. Therefore, it is an hybrid of a discrete event simulation engine and an expert system extended by a combinatorial optimization module and a communication module (for online applications). The RAO simulator can also be seen as a virtual machine for data-oriented programming.
By Barış Bozkurt (Izmir Institute of Technology (IYTE)) October 22nd, 2009
Automatic Transcription for Turkish Music
In this talk, first the pitch characteristics and notation of Turkish music will be presented. Then the rest of the talk will be dedicated to the signal processing algorithms needed for automatic transcription: pitch detection, makam recognition and tonic detection, pitch quantization and rhythm analysis. slides (pdf), short CV (doc)
By Thomas Dubuisson (TCTS Lab.) October 8th, 2009
Extraction and combination of acoustic descriptors for the detection of speech pathologies
This presentation aims at giving a brief overview of my PhD thesis. This thesis deals with the use of acoustic descriptors for the detection of speech pathologies. Two cases are presented: the use of descriptors from speech signals and the use of descriptors from both speech signal and glottal source estimation. The different descriptors are defined, as well as the methods involved in the different cases.
By Anderson Mills (TCTS Lab.) September 17th, 2009
Human-Based Percussion and Self-Similarity Detection in Electroacoustic Music
The goal of this research was to create an algorithm that identifies instants in pieces of music that humans would identify as percussive. The first step was to collect human judgments about percussive sounds in order to acquire sound-event dimensions that humans use as percussive cues. The next step was to design a percussivity-profile algorithm using a time-domain, channel-based approach and psychoacoustic models with input parameters tuned to maximally match the human choices. The final step was to use a similarity matrix combined with the percussivity-profile algorithm to find structural characteristics in the percussive sounds of electroacoustic music.
The results of the human data collection and of the percussivity-profile algorithm and similarity matrix will be presented for several examples.
By Maria Astrinaki (Computer Science Department, University of Crete, Greece) July 2nd, 2009
The aim of this short time scientific mission in TCTS Lab was to learn about Pure Data and real time programming, in order to use it for voice analysis and pathology detection. This presentation deals with the first results achieved.
Using Pure Data, the Normalized Noise Energy ratio, NNE, was implemented, in order to measure the noise corrupting the input signal. Also by using the autocorrelation method in the spectral domain a pitch estimator was implemented. Moreover there was an attempt for real time pitch synchronous signal analysis, which due to the limitations of Pure Data need further research. Pure Data was also used for Praat file oversampling, in order to have synchronicity with the audio input.
Furthermore, because of the not friendly interface of Pure Data for the visualization of the results, Processing was used.
By Prof. Mircea Mihaiu (Electronics and Instrumentation Department, University of Craiova, Romania) June 11th, 2009
- About the Electronix Modeling of Memristors :
  Memristors are a class of passive two-terminal circuit elements that maintain a functional relationship between the time integrals of current and voltage. we present the similarity and differences between memristor and phototransistor. We developed the new electronic models of the memristor based on known electronic devices such photoresistor and linear integrated circuits such AO. The memristor model can be used to understand the main functioning of the memristors and how to use such devices in some applications before the nanoscale memristor will be on the market. We present the main circuit, advantages and disadvantages and results after the simulation with MULTISIM software
- The Structure, Model and Simulation of Rectifier Circuits for RFID Tags :
  We present the analysis of the main structure of the MOS transistors used as MOS diodes or MOS switches in RF rectifiers for RFID tags. We present 5 different kinds of rectifier structures for comparison. They are: NMOS bridge rectifier structure, NMOS with two MOS gate cross-connected bridge rectifier structure, PMOS bridge rectifier structure, PMOS with two MOS gate cross-connected bridge rectifier structure, gate cross-connected PMOS and NMOS bridge rectifier. All the above structures are presented and analysed with the MULTISIM 10 software. The simulation results are presented. Also we propose the using of the new kind of zero threshold MOSFET transistors developed by ADL in rectifier structure.
By Daniel Tuyttens (Maths Lab.) May 7th, 2009
In this talk, a particular combinatorial optimization problem is discussed.
The cover printing problem consists in the grouping of book covers on offset plates in order to minimize the total production cost.
As the considered problem is hard to solve, a metaheuristic method is presented : a greedy random adaptive search procedure (GRASP). The obtained results are compared with those found in the literature.
By Fabrice Estievenart (CETIC) April 30th, 2009
Data extraction from Internet with Retroweb
To stay competitive, a company has to keep control over information related to itself. Information can be internal data as well as those published regularly on the Internet.
Although it is an invaluabale and inexhaustible informative source, Internet also presents characteristics that make it difficult and even inefficient to exploit (plentiful, unstructured and changing information)
In this speech I will introduce Retroweb, a tool intended to extract targeted data from the Internet. I will address the methodological, architectural and technical aspects of the software.
Finally I will demonstrate the tool in the framework of a study case on the extraction of discussion forum.
By Justine Decuypere (Electrical Engineering) April 2nd, 2009
Influence of Mesopic Lighting Conditions in Artificially Lighted Urban Environment
Mesopic conditions extend from sunny day conditions (photopic) to dark night conditions (scotopic).
Human retina is made up two sorts of photoreceptors: cones and rods. The one are responsible for photopic vision, in colours whereas the others allow scotopic vision, colourless. In mesopic conditions cones and rods work together.
Artificially lighted external scenes are typically mesopic. However, all photometric units (lux, candelas...) are weighted by the eye photopic spectral sensibility curve. That is why these units are inappropriate for light calculations at low light levels. This study emphasizes the need for new measure and calculation methods dedicated to external lighting
By Xavier Siebert (Maths Lab.) March 26th, 2009
MediaCycle - Browsing multimedia databases
The aim of this project is to develop a software for browsing multimedia libraries (sound, image, video), with potential applications as a performance tool.
The browsing is based on content features (e.g., rhythm / harmony / melody for sounds; color / shape / texture for images), as opposed to metadata (such as captions or keywords).
The current version of the software will be presented, and mathematical tools to assist navigation will be discussed.
By Damien Tardieu (TCTS Lab.) March 12th, 2009
Instrument Models for Computer Aided Orchestration
This presentation deals with the design of a new method for computer aided orchestration . Orchestration is seen as the art of manipulating the timbre of an orchestra by the combination of different instruments.
The orchestrating task we focus on is the following one : To find combinations of instrumental sounds whose timbre is close to a target sound provided by the composer. This target is specified by a recorded sound. The knowledge of the system on the instrument timbre is extracted from sample databases.
First, a set of features that describe the timbre of the instruments is presented. Then we propose a probabilistic model of these features. This model is based on a taxonomy of the various instruments playing techniques and on the approximation of the feature distributions by a product of Gaussians. The model parameters are learned on the sample databases. This model and the proposed learning technique allow the inference of sounds that are missing in the databases. From the instrument models, we find the model of a combination of sounds to assess the similarity between this combination and the target sound. Proposed methods are tested on solo and polyphonic sound classification tasks.
By Sullivan Hidot (Computer Science Lab.) February 19th, 2009
Statistical analysis of danced movement
By Céline Mancas-Thillou (Transfer Technology Office), February 12th, 2009
The Transfer Technology Office (TTO) of FPMs will present its main activities by detailing its missions and will explain its implication in a very large regional project in order to make valorization practices consistent in Wallonia. But the main objective of this seminar is first to meet all research teams of the IT cluster in order to get needs of every researcher. Moreover, it will enable the TTO to explain its own needs and future achievements to help researchers in their daily job life.
slides (ppt)
By Fabien Rogister (TCTS Lab.), February 5th, 2009
Anesthesia & sleep
By Jean-Yves Parfait (TCTS - Multitel), January 22nd, 2009
Bayesian filtering
By Jerome Meessen (Multitel), December 18th, 2008
Interactive Classification of visual surveillance scenes
In this talk, I will present and demonstrate an interactive system for retrieving video frames.
The approach consists in an iterative classification based on examples provided by the user, and integrates multiple-instance learning, active learning and an innovative user interface.
By Laurent Jojczyk (Electronics Lab.), November 20th, 2008
The Road-Network on Chip is a new communication architecture developed by the Semi1009. This architecture gives an alternative to traditional data bus when this solution becomes inappropriate(Multiprocessors systems, multicaptors,...). The Network on Chip solutions are able to improve the performances of the applications but for the moment, most of the designers are unfamiliar with this methodology. The objective of this presentation is to give a clear introduction on this subject and on how to integrate it to a project.
By Sébastien Noël (Computer Science Lab.), November 6th, 2008
Resource access negotiation in Grids : model and simulator
This talk will present my PhD research topic and will be divided in two parts.
The first one will show the interest of a peer-to-peer negotiation in a distributed system in order to obtain a decentralized infrastructure.
The second part will present a software tool developped to simulate such a negotiation network and to test different management policies for the sites in the Grid.
By Thomas Drugman (TCTS Lab.), October 16th, 2008
Modeling the glottal source in parametric speech synthesis
Statistical parametric speech synthesizers have recently shown their ability to produce natural sounding voices. They also gained considerable attention for their flexibility, smoothness and small footprint. Nevertheless their main disadvantage is the typical buzziness of the produced speech. This presentation addresses methods proposed to incorporate a more suited modeling of the source signal so as to enhance the delivered quality.
By Matei Mancas (TCTS Lab.), October 2nd, 2008
EyesWeb XMI
EyesWeb refers to an open software platform which supports the development of real-time multimodal distributed interactive applications. The EyesWeb project started in 1997, and the current release is EyesWeb XMI (eXtended Multimodal Interaction). We used this platform at eNTERFACE 2008 and found it to be promising. Its aim is to put together native EyesWeb blocs but also blocks using various signal processing and communication libraries with a free SDK. A demonstration of live data acquisition and processing will be achieved and some patch examples will be provided in order to show the interest of using this platform in image processing but not only...
By Kimcheng Kith (TCTS Lab.), June 26th, 2008
k-means algorithm
By Thierry Ravet (TCTS Lab.), June, 19th, 2008
Sleep stage automatic classification
By Géraldine Bous (Maths Lab.), June, 12th, 2008
Errare humanum est
In the context of decision support systems, information provided by the decisionmaker are generally assumed to be perfectly accurate and reliable. However it can be shown that the process associated with problems such as choice or storage, which are involved in the elicitation of the preferences are cognitively complex. Therefore it is necessary to consider the possibility that the information provided by the decisionmaker on its preferences are not error-free.
We consider some of the consequences and implications of this hypothesis on the problems of storage as well as the effectiveness and accuracy of related algorithms.
By Laurent Couvreur (TCTS Lab.), June, 5th, 2008
Processing
In this talk, we will give an overview of the Processing open source programming language and environment. Processing is basically a software sketchbook for people who want to program images, animation and interactions, and many other sweeties.
Relying on Java technology, it can be integrated in many other platforms and can be easily extended.Processing is ideally designed for students, artists, designers, researchers, and hobbyists for learning, prototyping, and production of digital arts, which makes it a perfect candidate for realizations in the framework of the NUMEDIART program of excellence.
By Jérôme Urbain (TCTS Lab.), May, 29th, 2008
Laughter
Laughter is an important element of human communication.
By Michel Bagein (Computer Science Lab.), May, 15th, 2008
How to pass from the design of an application to its implementation. Who of us was not perplexed vis-a-vis the development of a software? Very often, one easily imagines the algorithms which correspond to the awaited functionalities but one butts sometimes against their implementations homogeneous in a total application, and this particularly in the application to strong constraints which are the systems real-time. To systematize the design and the implementation of applications, this talk will make you a rapid tower of horizon of a methodology of design (SART) and development (DARTS) adapted to the systems real-time. These methods will be illustrated on a pragmatic example where it will be detailed for you each stage from the preliminary design of the system until its implementation.
slides (pdf)
By Christophe Caucheteur (Telecom Lab.), May, 8th, 2008
A Fiber Bragg grating (FBG) is a section of optical fiber whose core presents a periodic and permanent modulation of its refractive index.
By Mihaiu Mircea (Visiting prof., Electronics Lab.), April, 24th, 2008
We present the University of Craiova with all the faculties and specialization. There is a detailed presentation of the Automation, Computers and Electronics faculty. We also present the principles of RFID systems and our research to introducing RFID systems in production cycle for tire in Slatina factory Romania.
By Saïd Mahmoudi (Computer Science Lab.), April, 10th, 2008
In the first part of this presentation I will present my thesis work, in which I address the problem of 3D models retrieval and recognition using two dimensional shape representation of 3-D objects. I present a shape similarity system based on the correspondence of visual 2D parts. These parts are obtained by a shape segmentation approach using the Curvature Scale Space (CSS). I propose a 3D search engine based on these 3D models characteristic views. In the second part of the seminar I present our research activities in the computer science lab. The goal of our works is to propose and evaluate some new methods used to calculate vertebra orientation in medical X-ray images.
By Lionel Lawson (UCL), March, 20th, 2008
In this seminar we will demonstrate an open-source platform (OpenInterface), namely a set of tools aimed at rapid development and prototyping of multimodal applications. The platform tools are aimed at designing, implementing and testing multimodal interfaces that handle a rich and extensible set of modalities. It supports integration of heterogeneous (C/C++, Matlab, .NET, Java, etc.) components (devices, algorithms, applications, etc.) and quick replication, supports dynamic selection and combination of modalities to fit the ongoing context of use, and enables iterative user-centered design.
By Olivier Pietquin (Supelec-Metz), March, 12th, 2008
This seminar aims at presenting the activities of the "Information, Multimodality & Signal" (IMS) research group of the Supélec Metz Campus. This group combines competencies in signal processing and computer science applied to multimodal information processing. Especially, 3 research topics are investigated: statistical signal processing and machine learning, intensive and distributed computing, artificial intelligence and knowledge management. The pursued goal of the IMS research group is to design "situated systems" able to integrate perception, action and reasonning into a single process. The applicative domains chosen by the group are cognitive and interactive robotics as well as smart environments.
By Raphael Sebbe and Alexis Moinet (TCTS Lab.), March, 6th, 2008
Eurobot European cup of robotic - An overview of the Kadroïd team (Mons). This seminar presents the current activities of Kadroïd, an amateur robotic team composed of FPMs alumni, within the context of the Eurobot 2008 challenge. Team's background over past years will be presented as well as the rules of the 2008 challenge and the droïd global architecture. Technical details and choices of implementation at the mechanical, electronic and programming levels will be introduced as well. "What's on ?" : cutting-edge laser cutting for the aluminium structure, printed circuits board (PCB) design, finite state machines (FSM) programming and computer vision.
By Matthieu Aerssens (Electronics Lab.), February, 28th, 2008
The objective of this DEA work is the digital implementation of a decimator filter optimized for a IEEE 802.15.4 Zigbee receiver architecture. That work was enclosed within the European Medea+ project Witness (Crotale). The project objective is the development of new technologies for short-range communications using very low power while developing extensions to new and existing standards. Part of our activity was the design and implementation of an analog-digital converter consisting of a Sigma-Delta (Σ-Δ) modulator and associated decimator filter. The presentation will introduce the basis of Σ-Δ conversion (over-sampling, quantification noise shaping and decimation), the design flow details, and the optimization techniques involved in order to come up with a low power implementation. A full test scenario including a FPGA decimator filter prototype and the already integrated UMC 130nm Σ-Δ modulator was developed.
By Joelle Tilmanne (TCTS Lab.), February, 21st, 2008
The IGS-190 is a Motion Capture suit, that enables the user to record real-time 3D angles between the main body parts. We will see a quick demonstration of this system, what it can be used for and how to use it.
By Albdehak Lakhouaja (Computer Science Lab.), February, 7th, 2008
A morphological analyzer for Arabic words: in this presentation, we will present a system of morphological analyzis of the Arabic language, which enables to give all modifications of a word, given any Arabic word (without vowels).
By Mohamed Talibi Alaoui (Computer Science Lab.), Januray, 31st, 2008
Classification approach of color images by association of morphological transformations and Kohonen maps: we will present a new approach of unsupervised classification for color images, based on the use of auto-adaptive Kohonen maps and mathematical morphology.
By Michel Bagein (Computer Science Lab.), January, 24th, 2008
With one of the most efficient friction coefficient, the rail-wheel contact offers a very high economic level of power saving for railroads. In contrast, the very low adhesion capacity limits mechanical performances and involves slip on powered wheels that modify the traction and braking behaviour ... but not always in the bad way! How does sliding phenomena impact the position and speed estimation aboard train? This talk will offer an outline of problematic and present solutions for modelling real time simulation of adherence rail-wheel in the context of safety odometric embedded systems validation.
slides (pdf)
By Aurélie Servain (TCTS Lab.), January 17th, 2008
Shadows segmentation for silhouette extraction in the context of behaviour analysis or limb tracking. This presentation will be about the 6 months I have spent in Kingston University. First, I will introduce you the Digital Imaging Research Centre, then the differents shadows segmentation techniques.
By Mathieu Senelle (Maths Lab.), December, 20th, 2007
Numerical Simulation of a simulated moving bed (SMB) chromatographic process: Simulated mobile bed (SMB) chromatography is a well mastered separation technique in certain traditional sectors such as hydrocarbon and sugar separation. However, its application to the separation of high added value substance in pharmaceutical industry poses new problems related to the nature of the products to be separated, and also to more strict requirements concerning purity and produced quantities. A robust and efficient numerical simulator will be thus a valuable help to separate in an optimal way these compounds, for instance by representing the changes induced by a variation of the installation parameters. This presentation will include an introduction to chromatography and to the SMB process, and will then focus on numerical aspects which make possible to simulate this process.
By Sébastien Noël (Computer Science Lab.), November, 29th, 2007
Resource access negotiation in grid infrastructures. During this talk, grid computing infrastructures will be presented and detailed in their different software components. Difficulties induced by commercial usage of such infrastructures will be underlined and will lead to the description of my research work: the creation of management tools for resource access negotiation in grids.
By Valéry Tamgnoue (Telecom Lab.), November, 22nd, 2007
Hierarchical modulation allows two separate bit streams to be modulated on one single stream according to their relative priority. Indeed, if two bit streams need to be transmitted in hierarchical modulation, primary constellation carries highest priority bit stream, while secondary constellation bears lowest priority bit stream. Hierarchical modulation is now approved in DVB-{T, H} (Digital Video Broadcasting - Terrestrial and Handheld) where it opens new future to mobile and portable television. We will broadly present the technical principles of hierarchical modulation: what is it? How does it work? Where is it used? And we will explain challenges faced by hierarchical modulation.
By Jonathan Demeyer (TCTS Lab.), November, 8th, 2007
My presentation will focus on vocal fold segmentation and methods that will achieve it. An overview of the state-of the-art of my research will be made including active contour models. Then, I will present an algorithm I'm currently working on : the Active Shape Models.
By Tinne Tuytelaars (KUL), October, 25th, 2007
Tinne Tuytelaars, researcher at KUL, in the ESAT-PSI-Visics lab, will present a brief tour through her lab with description of main image processing activities, followed by a more in-depth discussion on local feature extraction and object recognition.
By Thomas Dubuisson (TCTS Lab.), October, 18th, 2007
The talk will start with the presentation of the ECLIPSE project in terms of implicated partners and objectives. The estimation of glottal source being an important part of this project, the talk will continue with a brief description of the methods proposed in litterature for then focusing on the method based on the ZZT (Zeros of the Z-Tranform) representation. Finally application and improvement of this decomposition method will be presented for real speech signals.
By Thomas Drugman (TCTS Lab.), October, 11th, 2007
The Lombard effect: analysis and applications: the Lombard effect refers to the speech changes due to the immersion of the speaker in a noisy environment. These modifications are observed on an acoustic, phonetic as well as an articulatory point of view. Through an hyper-articulation (unconsciously most of the time), the speaker placed in a communicative context aims at maximizing the intelligibility of his utterances. After an analysis of the different changes produced, hindrances induced in automatic speech recognition and future potential applications in speech synthesis will be discussed.
By Gautier Ravet (Telecom Lab.), October, 4th, 2007
Optical time domain reflectometry for the distributed measurement of Raman fiber amplifiers: the goal of this presentation is to give an insight about the research lead in the SET in the field of nonlinear optics. The development of a new measurement method based on optical time domain reflectometry allowing the characterization of optical fiber amplifiers based on the Raman effect, a non linear effect, will be presented. After a brief introduction to the concept of Raman amplification and to the reflectometric techniques, we will present from the theoretical point of view the method that we elaborated. Then we will show the practical realization of the system and give the main results that we obtained with a suitable experimental data processing. We will finally demonstrate how to apply this method to the characterization of the four-wave mixing, another nonlinear effect, and its application to the measurement of the chromatic dispersion.
By Nicolas D'Alessandro (TCTS Lab.), September, 20th, 2007
HandSketch 1.x, Digital Instrument for Singing Synthesis: in the last talk, different issues of digital luthery have been exposed. Once this context is defined, the following talk will go a little bit further in my PhD thesis investigations. First some milestones of the way expressivity is defined -from sound and gestural points of view- will be presented. Then, by going further into details of singing voice synthesis, some of the underlying aspects of my work will be gathered and discussed around a concrete projet, HandSketch 1.x, presented at NIME conference this year. A demonstration will be performed and the instrument will be available for curious people.
By Pablo Godoy (TCTS Lab.), September, 13th, 2007
The talk will start by introducing the F3M project and its context. We will review the kind of services the project should provide and how to realise it. Then we will discuss about X+V language and vocal browser. Then we will look at our work which started with the XFL language and its interpreter : « Nimble ».We will go deeper int the implementation of Nimble and the discussion will take end with the examination of real applications and graphcet for Nimble.
By Sébastien Bette (Telecom Lab.), September, 6th, 2007
Research activities in the Electromagnetism and Telecommunications department
The two main research areas in which the Electromagnetism and Telecommunications department evolves are the characterisation of the physical layer of telecom systems and the optical fibre. In both cases, the metrology is an important activity. During the talk, we will globally present the different studies related to these two researches areas and their main issues. In addition, the department possesses fibre Bragg grating manufacturing facility. We will describe the features of theses optical components as well as different applications for telecommunication and sensing purposes.
By Jean-François Bauwens (Computer Science Lab.), June, 28th, 2007
Assessment of segmentation methods for medical imaging.
By Thierry Dutoit (TCTS Lab.), June, 21st, 2007
Presentation of the NUMEDIART project.
By Céline Mancas-Thillou (TCTS Lab.), June, 7th, 2007
Presentation of the RECITE project: what can one read ?
By Laurent Couvreur (TCTS Lab.), May, 31st, 2007
Presentation of the project entitled SERKET (SEcuRity KEeps Threats away) : open platform for multimodal surveillance.
By Raphaël Sebbe (TCTS Lab.), May, 24th, 2007
Presentation of eX-cinder activities.
By Clovis Dongmo Jiogo (Computer Science Lab.), May, 5th, 2007
Visualization and Virtual Reality on computer grids.
By Thibaut Lust (Maths Lab.), April, 26th, 2007
Metaheuristics for combinatorial and multicriteria optimization.
By Anna Auria, an Erasmus student (TCTS Lab.), April, 19th, 2007
Presentation of end-studies work entitled "An HMM-based speech synthesizer in French".
By Matthieu Jottrand (TCTS Lab.), April, 5th, 2007
Presentation of the conference paper: "Towards Voice Conversion System Based on Frame Selection" written by T. Dutoit, A. Holzapfel, M.Jottrand, A. Moinet, J.Pérez and Y. Stylianou.
By Frédéric Bettens (TCTS Lab.), March, 8th, 2007
Presentation of the project entitled MOUSTIC (« MObile USers Traffic Information Channels »).
By Nicolas D'Alessandro (TCTS Lab.), March, 1st, 2007
Digital Luthery : an interdisciplinary approach: Application to the challlenge of « vocal instrument ».
By Alexis Moinet (TCTS Lab.), February, 22nd, 2007
Presentation of the IRMA project.
By Stéphanie Devuyst (TCTS Lab.), February, 15th, 2007
Automatic classification into sleep stages: feature extraction and comparison of main classifiers.
By Silvio Ferreira (TCTS Lab.), February, 1st, 2007
Presentation of the Sypole project.

Faculté Polytechnique de Mons
9, rue de Houdain, B-7000 Mons, Belgium
Tél: +32 (0)65 374111 Fax: +32 (0)65 374200