The image segmentation based on transforming RGB color space and the shape classifier based on signature
feature are used to detect traffic signs in urban scenes. For improving recognition accuracy, two modal representations are
presented to classify the traffic sign. 1) The feature of traffic sign is extracted by dual-tree complex wavelet transform(DT-
CWT) and 2D independent component analysis(2DICA), then sent to the nearest neighbor classifier to classify traffic sign.
2) The template matching based on intra pictograms of traffic sign is applied to recognition. The recognition results are fused
by some decision rules. Experimental results show that, the overall recognition rate of the proposed algorithm is more than
91% and the average frame rate is up to 6.6fps, which indicates that the system is robust, effective and accurate to classify
traffic signs.