Traffic sign detection based on voting scheme and link distribution model

In this study, we propose a novel shape-based traffic sign detection method which consists of two stages. First, a rotational symmetry voting scheme is proposed to detect the centers and boundary sets of the candidate polygons in the image. Second, a Link Distribution (LD) model, which considers a polygon as the collection of links between center and boundary points, is proposed to refine the detection results and classify the polygons into the specific categories. We learn the LD models of triangle, diamond, octagon and circle using training samples. The candidate center and boundary points introduced by the voting scheme are matched with all the four types of the LD models. The order of the links remains invariant even under deformation; therefore, the LD model can detect polygons with shape deformation and partial occlusion. Experimental results show that our algorithm outperforms several existing road sign detection methods with higher precision, better generality and real-time capability when evaluated using the public dataset.