Detection of traffic signs based on visual saliency model

In this paper, a visual saliency model is used to detect the traffic signs in the natural images. A number of features consist of intensity, color opponency and orientations are extracted from the image by the model. Without the need for image segmentation, the saliency map indicating the area where the traffic signs are can be computed. From the experiments on the real scene images with the domestic traffic signs, the detection rate of 89% is achieved.