This paper presents a method to create new images of a scene from a small set of calibrated images using color matching. While most of the existing techniques attempt to build a 3d reconstruction of the scene, the only geometric constraint our color matching method requires is camera calibration. This paper describes a new adaptation of the plane-sweeping algorithm using a local dynamic scoring...

This project is a part of a bigger project aiming to develop bio-inspired climbing robots for space exploration. My goal is to provide a tool to simulate the physics of those robots, design them and test their controllers. Therefore my project is divided into three main steps. The first step consists in developing a physical plugin for Webots in order to be able to simulate the bio-inspired adh...

The IRCAM Musical Acoustics Team is not only focused on the musical instruments modelisation research but also on the computer implementation of these models and the evaluation of the musical interest that the computer implementation may have. The physical modelling sound synthesis software Modalys is an environment where virtual acoustical instruments may be imagined, performed and manipulated...

In this paper, we introduce a model for Discrimant Feature Extraction (DFE): the Neural Predictive Coding (NPC). It is an extension of the Linear Predictive Coding (LPC). The Modelisation Error Ratio (MER), a discriminant criterion adapted for predictive models, is introduced. We propose a theoretical validation of the discriminant properties of the MER. The experimental validation consists on ...

In this paper, we describe a new approach, allowing as much as possible, an automatic extraction of the micromelodic information of the speech signal using the original curved of the fundamental frequency and its macromelodic curve obtained by MOMEL, the algorithm of MOdelisation MELodic. The results reinforce the idea that microprosodic effect exists and seems to be possible to be included in ...

Multiingered robot hands can be used for applying ne motions. Unlike pick-and-place manipulations, Fine Motions Strategies generate movements based on sensor information to overcome mechanical and sensory uncertainties(Laugier, 1989; Magnussen and Doersam, 1995; Munoz et al., 1995). Performing assembly operations with a dexterous hand, high level strategies for planning a manipulation task have...