Titre : A Vision System for Segmentation and Analysis of Colored Cytological Images

Auteur(s) : F. HACHOUF, S. OUADFEL **

Laboratoire(s) : Université de Constantine, Département d'électronique,
Route Ain-El-Bey, 25000 ALGERIE
Fax : 213-31-61-42-06, Tel:213-31-61-42-07 poste 36.

Résumé :

In this paper, we present a vision system to segment and analyze colored cytological images. The adopted approach combines, at a time, the principles of region growing and edge detection guided by a fuzzy classification. So the presented cells are presegmented using the fuzzy C-means algorithm, followed by a hybrid segmentation algorithm. Measures of shapes and colors are taken for every cellular object for analysis and interpretation purposes.

Liste de mots-clefs :

Segmentation - Analysis- cytological images - region-growing edge detection - fuzzy classification