Robust Japanese Road Sign Detection and Recognition in Complex Scenes Using Convolutional Neural Networks

In recent years, the development of image processing technologies used for autonomous driving has been remarkable. Automatic detection and recognition of road signs are required for the practical use of autonomous vehicles. In the detection and recognition of road signs, changes in scale and contrast greatly affect the accuracy. In this study, we solve this problem by learning road signs using a deep learning technique that is robust against scale changes, and thought an experiment, we compare our method with recently proposed deep learning methods. We also show the results using our proposed method for individual Japanese road signs. The proposed method shows higher accuracy in the detection and recognition of road signs than the faster Region-based Convolutional Neural Network (Faster R-CNN) and Single Shot multibox Detector