Dans cet article, nous proposons une nouvelle approche d'analyse de forme pour la détection des rétinopathies vasculaires. La méthode procède à la classification des vaisseaux et quantifie la modification de leur tortuosité sur les images d'angiographie rétinienne en fluorescence. La classification est basée sur une technique de division/fusion. Deux classes sont définies: artère/veine et artériole/veinule. Dans une séquence d'images, les vaisseaux classifiés sont mis en correspondance. La quantification consiste à comparer la tortuosité des vaisseaux classifiés correspondants. La tortuosité est définie par la valeur de l'excentricité déterminée avec les moments invariants 2D d'ordre 2. La méthode est appliquée aux images d'angiographie rétinienne en fluorescence de patients diabétiques et drépanocytaires.
The reduction of the blur and the noise is an important task in image processing. Indeed, these two types of degradation are some undesirable components during some high level treatments. In this paper, we propose an optimization method based on neural network model for the regularized image restoration. We used in this application a modified Hopfield neural network. We propose two algorithms using the modified Hopfield neural network with two updating modes : the algorithm with a sequential updates and the algorithm with the n-simultaneous updates. The quality of the obtained result attests the efficiency of the proposed method when applied on several images degraded with blur and noise.
Although human face recognition is a hard topic due to many parameters involved (e.g. variability of the position, lighting, hairstyle, existence of glasses, beard, moustaches, wrinkles...), it becomes of increasing interest in numerous application fields (personal identification, video watch, man machine interfaces...). In this work, we present WHO_IS, a system for person identification based on face recognition. A geometric model of the face is definedfrom a set of characteristic points which are extracted from the face image. The identification consists in calculating the K nearest neighbors of the individual test by using the City-Block distance. The system is tested on a sample of 100 people with a success rate of 86 %.
One characteristic of planar topological graphs is the reversal degree. In this paper, we propose an improve algorithm for calculating the reversal degree of a planar topological graphs. This algorithm explores various possible cases following the descending method. Practical tests carried out on machine, using graphs with more than fifty internal vertices of odd degree, have been realized within reasonable computing time.
Tools dedicated to mathematics need to display formulas and to interact with them. In this paper, we present a summary of existing tools, then we describe FIGUE, an incremental two dimensional layout engine, developed at INRIA, to get a specialized toolbox for building customized editors and graphical user interfaces. Finally we give an exemple of interface using FIGUE to develop mathematical proofs on computer.
Inverse modeling has become a standard technique for estimating hydrogeologic parameters. These parameters are usually inferred by minimizing the sum of the squared differences between the observed system state and the one calculed by a mathematical model. Since some hydrodynamics parameters in Richards model cannot be measured, they have to be tuned with respect to the observation and the output of the model. Optimal parameters are found by minimizing cost function and the unconstrained minimization algorithm of the quasi-Newton limited memory type is used. The inverse model allows computation of optimal scale parameters and model sensi-tivity.
