ROBUST COLOR OBJECT DETECTION AND RECOGNITION (WedPmPO4)

Author(s) : Maaoui Choubeila (Ecole Nationale Supérieure d'Ingénieurs de Bourges, France) Christophe Rosenberger (LVR, France) Sébastien Chabrier (LVR, France)

Abstract : In this paper, we propose a new method for the detection and the recognition of objects in color images. The developed method processes in two steps. The first one concerns the object detection by a color segmentation method and a labeling algorithm. The second step deals with the recognition of extracted objects. Each object is described by color Zernike moments. In order to recognize each object, we use a support vector machine. We illustrate the efficiency of the proposed method on color images embedding some objects from the COIL-100 database.

Menu