Exploitation de la couleur pour la segmentation et l'analyse d'images. (Color image segmentation)

In this paper, we propose an original approach in order to improve the results of color im-age segmentation by pixel classification. We define a new kind of color space by selecting a setof color components which can belong to any of the different classical color spaces. Suchspaces, which have neither psycho-visual nor physical color significance, are named hybridcolor spaces. We propose to classify pixels represented in the hybrid color space which is spe-cifically designed to yield the best discrimination between the pixel classes. This space, which iscalled the adapted hybrid color space, is built by means of a sequential supervised feature se-lection scheme. This procedure determines the adapted hybrid color space associated with agiven family of images. Its dimension is not always equal to three, as for classical color spaces.The effectiveness of our color segmentation method is assessed in the framework of soccer im-age analysis. The team of each player is identified by the colors of its soccer suit. The aim ofthe segmentation procedure is to extract meaningful regions representing the players and torecognize their teams.2003 Elsevier Science (USA). All rights reserved.