Road sign detection by clustering the EMD based interest points

This paper addresses an important issue - automatic road traffic signs detection and recognition in natural scenes, which is accomplished in the three following stages: interest points detection, clustering of those points and similarity search. At the first stage, good discriminative, rotation and scale invariant interest points are selected from the image edges based on the 1-D empirical mode decomposition (EMD). Then stable local features related to the brightness and color are extracted using Gabor filter and the detected points are clustered to find the possible candidate road signs or the region of interests (ROIs). We propose a two-step unsupervised clustering technique, which is adaptive. Finally, a fringe-adjusted joint transform correlation (JTC) technique is used for matching the unknown signs with the existing known reference road signs in the database. The presented framework provides a novel way to retrieve a road sign from the natural scenes.