In this paper, we present a color segmentation method based on a normalized color-naming algorithm, which removes the effects of the varying conditions due to changes in scene illuminant. Images labeled with the color name and intensity of small regions is further processed by a region-growing step providing a sound segmentation. The method has been tested on a large set of images we get from a surveillance system, whose goal is the automatic retrieval of people from an image database using their appearance description. It is given in terms of placement of the color regions in clothes. Finally, a quantitative measurement to evaluate the performance of the algorithm has been defined.