{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/gs3d-an-efficient-3d-object-detection","title":"GS3D: An Efficient 3D Object Detection Framework for Autonomous Driving","arxiv_id":"1903.10955","date":"2019-03-26","proceeding":"CVPR 2019 6","authors":["Buyu Li","Wanli Ouyang","Lu Sheng","Xingyu Zeng","Xiaogang Wang"],"abstract":"We present an efficient 3D object detection framework based on a single RGB\nimage in the scenario of autonomous driving. Our efforts are put on extracting\nthe underlying 3D information in a 2D image and determining the accurate 3D\nbounding box of the object without point cloud or stereo data. Leveraging the\noff-the-shelf 2D object detector, we propose an artful approach to efficiently\nobtain a coarse cuboid for each predicted 2D box. The coarse cuboid has enough\naccuracy to guide us to determine the 3D box of the object by refinement. In\ncontrast to previous state-of-the-art methods that only use the features\nextracted from the 2D bounding box for box refinement, we explore the 3D\nstructure information of the object by employing the visual features of visible\nsurfaces. The new features from surfaces are utilized to eliminate the problem\nof representation ambiguity brought by only using a 2D bounding box. Moreover,\nwe investigate different methods of 3D box refinement and discover that a\nclassification formulation with quality aware loss has much better performance\nthan regression. Evaluated on the KITTI benchmark, our approach outperforms\ncurrent state-of-the-art methods for single RGB image based 3D object\ndetection.","url_abs":"http://arxiv.org/abs/1903.10955v2","url_pdf":"http://arxiv.org/pdf/1903.10955v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[],"tasks":[{"task_slug":"3d-object-detection","task_name":"3D Object Detection"},{"task_slug":"autonomous-driving","task_name":"Autonomous Driving"},{"task_slug":"monocular-3d-object-detection","task_name":"Monocular 3D Object Detection"},{"task_slug":"object","task_name":"Object"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"vehicle-pose-estimation","task_name":"Vehicle Pose Estimation"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/monocular-3d-object-detection-on-kitti-cars","task":"Monocular 3D Object Detection","dataset":"KITTI Cars Moderate","model":"GS3D","rank_in_archive_order":28,"of":29,"metrics":{"AP Medium":"2.9"},"uses_additional_data":false},{"leaderboard":"/sota/vehicle-pose-estimation-on-kitti-cars-hard","task":"Vehicle Pose Estimation","dataset":"KITTI Cars Hard","model":"GS3D","rank_in_archive_order":18,"of":19,"metrics":{"Average Orientation Similarity":"61.85"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1903.10955","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}