Traffic sign detection is an important task in traffic sign recognition systems. Chinese traffic\nsigns have their unique features compared with traffic signs of other countries. Convolutional neural\nnetworks (CNNs) have achieved a breakthrough in computer vision tasks and made great success in\ntraffic sign classification. In this paper, we present a Chinese traffic sign detection algorithm based\non a deep convolutional network. To achieve real-time Chinese traffic sign detection, we propose\nan end-to-end convolutional network inspired by YOLOv2. In view of the characteristics of traffic\nsigns, we take the multiple 1 Ã?â?? 1 convolutional layers in intermediate layers of the network and\ndecrease the convolutional layers in top layers to reduce the computational complexity. For effectively\ndetecting small traffic signs, we divide the input images into dense grids to obtain finer feature maps.\nMoreover, we expand the Chinese traffic sign dataset (CTSD) and improve the marker information,\nwhich is available online. All experimental results evaluated according to our expanded CTSD and\nGerman Traffic Sign Detection Benchmark (GTSDB) indicate that the proposed method is the faster\nand more robust. The fastest detection speed achieved was 0.017 s per image.
Loading....