Image segmentation based road sign detection

This paper proposes an automatic method to detect road traffic signs in natural scenes. There are three main stages in the proposed algorithm: (1) segmentation based on the brightness and color features to find the possible candidate road sign regions; (2) sign detection by using two shape classification criteria; and (3) recognition of the road sign by employing a fringe-adjusted joint transform correlation (FJTC) technique. The proposed framework provides a novel way to detect a road sign by integrating image features with the geometric shape information. Experimental results on real life images demonstrate that the proposed algorithm is invariant to translation, rotation, and scale.