{"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/lmnet-real-time-multiclass-object-detection","title":"LMNet: Real-time Multiclass Object Detection on CPU using 3D LiDAR","arxiv_id":"1805.04902","date":"2018-05-13","proceeding":null,"authors":["Kazuki Minemura","Hengfui Liau","Abraham Monrroy","Shinpei Kato"],"abstract":"This paper describes an optimized single-stage deep convolutional neural\nnetwork to detect objects in urban environments, using nothing more than point\ncloud data. This feature enables our method to work regardless the time of the\nday and the lighting conditions.The proposed network structure employs dilated\nconvolutions to gradually increase the perceptive field as depth increases,\nthis helps to reduce the computation time by about 30%. The network input\nconsists of five perspective representations of the unorganized point cloud\ndata. The network outputs an objectness map and the bounding box offset values\nfor each point. Our experiments showed that using reflection, range, and the\nposition on each of the three axes helped to improve the location and\norientation of the output bounding box. We carried out quantitative evaluations\nwith the help of the KITTI dataset evaluation server. It achieved the fastest\nprocessing speed among the other contenders, making it suitable for real-time\napplications. We implemented and tested it on a real vehicle with a Velodyne\nHDL-64 mounted on top of it. We achieved execution times as fast as 50 FPS\nusing desktop GPUs, and up to 10 FPS on a single Intel Core i5 CPU. The deploy\nimplementation is open-sourced and it can be found as a feature branch inside\nthe autonomous driving framework Autoware. Code is available at:\nhttps://github.com/CPFL/Autoware/tree/feature/cnn_lidar_detection","url_abs":"http://arxiv.org/abs/1805.04902v2","url_pdf":"http://arxiv.org/pdf/1805.04902v2.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":[{"paper_slug":"lmnet-real-time-multiclass-object-detection","repo_url":"https://github.com/CPFL/Autoware","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"autonomous-driving","task_name":"Autonomous Driving"},{"task_slug":null,"task_name":"CPU"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}