Traffic sign detection and recognition via transfer learning

Automatic driving has become a extremely hot issue in recent years, and the detection of stop signs is critical for autonomous driving. Different from precious methods in which target features were extracted and then feed to SVM classifier to classify different types of traffic signs, this paper introduces a kind of transfer learning method based on the convolutional neural network(CNN). A deep convolution neural network is trained using a large data sets, and then a valid region convolutional neural network(RCNN) detection can be obtained through a small amount of traffic standard training samples. At the end of this paper, the classic GTSDB data sets and some other data of shenzhen university town are used to show the effectiveness of the transfer learning approach.