Traffic sign recognition using hybrid features descriptor and artificial neural network classifier

Abstract

Traffic Sign Recognition (TSR) system is a significant component of Intelligent Transport System (ITS) as traffic signs assist the drivers to drive more safely and efficiently. This paper represents a new approach for TSR system using hybrid features formed by two robust features descriptors, named Histogram Oriented Gradient(HOG) features and Speeded Up Robust Features(SURF) and artificial neural network (ANN) classifier. In the detection step, the region of interest (sign area) is segmented using color based thresholding algorithm, post processed to filter the unwanted region. Next robust features vector named Distance to Borders (DtBs) of the segmented blob is formed to verify the shape of the traffic sign. Finally the recognition of the traffic sign is implemented using ANN classifier upon the training of hybrid features descriptor. The proposed system simulated on offline road scene images shows a high classification rate in the recognition stage. The performance of the ANN model is illustrated in terms of cross entropy, confusion matrix and receiver operating characteristic (ROC) curves. In addition, the performance of hybrid feature descriptor is compared with recognition based on HOG and SURF descriptor respectively. Also, performances of some classifier such as Support Vector Machine (SVM), Decision Trees, Ensembles Learners (Adaboost) and K-Nearest Neighbor (KNN) classifier are assessed with ANN approach. The simulation results illustrates that recognition using hybrid feature descriptor outperforms in all classifier and the recognition accuracy of ANN is higher than classifier stated above.