The subject of this paper is on drtection, location of traffic signs. The way we are operating is by first detecting salient features, then tracking them, and finally identifying then1 in a snapshot view at the right resolution. A mixed model guided active detection is chosen to compensate for imperfect low level segmentation. The detection is also optimized based on a active contour model. The first experimental results are presented and the future working directions on the topic are also discussed.