2017/10/17 by Seokju Lee, Junsik Kim, Lee, Seokju +17 · 2 citations
Computer Science · Engineering · #Advanced Neural Network Applications #Autonomous Vehicle Technology and Safety #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Image and Object Detection Techniques
paper · pdf · doi:10.48550/arxiv.1710.06288
openalex publication_date 2017/10/17 · openalex created_date 2022/10/04 · openalex updated_date 2026/07/28
In this paper, we propose a unified end-to-end trainable multi-task network\nthat jointly handles lane and road marking detection and recognition that is\nguided by a vanishing point under adverse weather conditions. We tackle rainy\nand low illumination conditions, which have not been extensively studied until\nnow due to clear challenges. For example, images taken under rainy days are\nsubject to low illumination, while wet roads cause light reflection and distort\nthe appearance of lane and road markings. At night, color distortion occurs\nunder limited illumination. As a result, no benchmark dataset exists and only a\nfew developed algorithms work under poor weather conditions. To address this\nshortcoming, we build up a lane and road marking benchmark which consists of\nabout 20,000 images with 17 lane and road marking classes under four different\nscenarios: no rain, rain, heavy rain, and night. We train and evaluate several\nversions of the proposed multi-task network and validate the importance of each\ntask. The resulting approach, VPGNet, can detect and classify lanes and road\nmarkings, and predict a vanishing point with a single forward pass.\nExperimental results show that our approach achieves high accuracy and\nrobustness under various conditions in real-time (20 fps). The benchmark and\nthe VPGNet model will be publicly available.\n