Construction de caractéristiques par programmation génétique pour un système de reconnaissance multiclasse

Bourgoin, Brice (2007). Construction de caractéristiques par programmation génétique pour un système de reconnaissance multiclasse. Mémoire de maîtrise électronique, Montréal, École de technologie supérieure.

Prévisualisation

PDF Télécharger (5MB) | Prévisualisation

Résumé

L'objectif est d'optimiser la reconnaissance automatisée d'objets en vision ou en télédétection. Nous partons du constat que les classificateurs sont sensibles à l'espace de représentation des données et qu'un remaniement de cet espace pourrait aider certains d'entre eux. Nous poursuivrons deux objectifs: avoir un concept de système qui nécessitera le moins possible d'interventions humaines lors de sa mise en place et minimiser le taux d'erreurs absolu.

Pour cela, nous utiliserons un algorithme de Programmation Génétique avec coévolution. Son objectif sera de construire un nouvel ensemble de caractéristiques en se basant sur son potentiel de classification selon les plus proches voisins. Ce jeu de caractéristiques sera testé sur différents classificateurs. Nous avons alors observé que tous les classificateurs avaient de bonnes performances sur le nouvel espace de représentation. Aussi, les performances absolues semblent avoir été améliorées, en particulier avec l'usage d'une machine à vecteur de support en aval.

Titre traduit

Feature construction by genetic programming for a multiclass recognition system

Résumé traduit

The goal of this research is to optimize the automated recognition of objects in computer vision or remote sensing applications. The premise is that classifiers are sensitive to the data representation space and that a reorganization of this space could improve the performance of some of them. We aim at two goals: to propose a framework for a system requiring the !east possible human interventions at the time of its installation and to minimize the absolute error rate.

For that, we have used a Genetic Programming algorithm with coevolution. Its objective was to build a new set of characteristics being based on its potential for classification according to its closest neighbours. Then this set of characteristics was tested on several types of classifiers: closer neighbours, artificial neurons networks and support vector machines.

In order to better target his research, we preferred to restrict the Genetic Programming algorithm at the reorganization of the representation space than to generate a complete classifier. Thus, we hoped to benefit from the force of advanced classifier such as support vector machines to prevent the Genetic Programming algorithm from reinventing what is already known. The algorithm had for only objective to concentrate on what is sometimes a weakness in a classification system: the data representation
space.

We have used two completely distinct data bases: the first containing handwritten digits, the second concerned with the differentiation of cereals such as barley, corn or oats. The first base contains ten classes, the second seven. Thus, they are real problems of computer vision and strongly multiclass systems. In addition to confirming the results, the interest in using two bases was to highlight the reduced need for human interventions in the initial setup of a classification system. Indeed, for the second base, we have used exactly the same parameters as those selected for the first: these internai parameters of our algorithm claim to be rather universal.

Several simulations allowed us to observe good performances with new space representations for whatever final classifier we used. In that, the robustness of the proposed system in reorganizing the representation space seems to offer improved performance when compared to a single classifier. Thus, we demonstrated the possibility of reducing human interventions needed for the system installation.

Moreover, the absolute performances seem to be improved, in particular with the use of a support vector machine downstream. This improvement was not always huge, but seemed sufficiently promising to pursue our investigation further. lndeed, our approach still offers many ways for improvement, mainly possible thanks to the many possibilities offered by algorithms based on the Genetic Programming paradigm.

Type de document:	Mémoire ou thèse (Mémoire de maîtrise électronique)
Renseignements supplémentaires:	"Mémoire présenté à l'École de technologie supérieure comme exigence partielle à l'obtention de la maîtrise en génie de la production automatisée." Bibliogr. : f. [140]-147.
Mots-clés libres:	Automatise, Caracteristique, Construction, Erreur, Espace, Genetique, Multiclasse, Optimisation, Programmation, Reconnaissance, Representation, Systeme, Teledetection, Vision
Directeur de mémoire/thèse:	Directeur de mémoire/thèse Landry, Jacques-André
Programme:	Maîtrise en ingénierie > Génie de la production automatisée
Date de dépôt:	06 avr. 2011 16:26
Dernière modification:	09 nov. 2016 02:11
URI:	https://espace.etsmtl.ca/id/eprint/561

Gestion Actions (Identification requise)

Dernière vérification avant le dépôt