{"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/super-accurate-low-latency-object-detection","title":"Super accurate low latency object detection on a surveillance UAV","arxiv_id":"1904.02024","date":"2019-04-03","proceeding":null,"authors":["Maarten Vandersteegen","Kristof Vanbeeck","Toon goedeme"],"abstract":"Drones have proven to be useful in many industry segments such as security\nand surveillance, where e.g. on-board real-time object tracking is a necessity\nfor autonomous flying guards. Tracking and following suspicious objects is\ntherefore required in real-time on limited hardware. With an object detector in\nthe loop, low latency becomes extremely important. In this paper, we propose a\nsolution to make object detection for UAVs both fast and super accurate. We\npropose a multi-dataset learning strategy yielding top eye-sky object detection\naccuracy. Our model generalizes well on unseen data and can cope with different\nflying heights, optically zoomed-in shots and different viewing angles. We\napply optimization steps such that we achieve minimal latency on embedded\non-board hardware by fusing layers, quantizing calculations to 16-bit floats\nand 8-bit integers, with negligible loss in accuracy. We validate on NVIDIA's\nJetson TX2 and Jetson Xavier platforms where we achieve a speed-wise\nperformance boost of more than 10x.","url_abs":"http://arxiv.org/abs/1904.02024v1","url_pdf":"http://arxiv.org/pdf/1904.02024v1.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":"super-accurate-low-latency-object-detection","repo_url":"https://gitlab.com/EAVISE/jetnet","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"object","task_name":"Object"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"object-tracking","task_name":"Object Tracking"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}