On-road object detection using deep neural network

Abstract

Industrialization of transportation system has derived serious accidents that resulted in thousands of deaths. To solve the problem, vision based object detection for autonomous vehicle and advanced driver assistance system has been researched. In this study, we provide experimentations of object detection and localization in on-road environment using deep neural network. We compared the detection accuracy among object classes and analyzed the recognition results with fine-tuned Single shot multibox detector on KITTI dataset. This work improves the performance of original detection model by increasing precision of overall detection about 6%, especially about 10% in pedestrian and cyclist.