{"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/a-hierarchical-deep-architecture-and-mini","title":"A Hierarchical Deep Architecture and Mini-Batch Selection Method For Joint Traffic Sign and Light Detection","arxiv_id":"1806.07987","date":"2018-06-20","proceeding":null,"authors":["Alex D. Pon","Oles Andrienko","Ali Harakeh","Steven L. Waslander"],"abstract":"Traffic light and sign detectors on autonomous cars are integral for road\nscene perception. The literature is abundant with deep learning networks that\ndetect either lights or signs, not both, which makes them unsuitable for\nreal-life deployment due to the limited graphics processing unit (GPU) memory\nand power available on embedded systems. The root cause of this issue is that\nno public dataset contains both traffic light and sign labels, which leads to\ndifficulties in developing a joint detection framework. We present a deep\nhierarchical architecture in conjunction with a mini-batch proposal selection\nmechanism that allows a network to detect both traffic lights and signs from\ntraining on separate traffic light and sign datasets. Our method solves the\noverlapping issue where instances from one dataset are not labelled in the\nother dataset. We are the first to present a network that performs joint\ndetection on traffic lights and signs. We measure our network on the\nTsinghua-Tencent 100K benchmark for traffic sign detection and the Bosch Small\nTraffic Lights benchmark for traffic light detection and show it outperforms\nthe existing Bosch Small Traffic light state-of-the-art method. We focus on\nautonomous car deployment and show our network is more suitable than others\nbecause of its low memory footprint and real-time image processing time.\nQualitative results can be viewed at https://youtu.be/_YmogPzBXOw","url_abs":"http://arxiv.org/abs/1806.07987v2","url_pdf":"http://arxiv.org/pdf/1806.07987v2.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":"a-hierarchical-deep-architecture-and-mini","repo_url":"https://github.com/bosch-ros-pkg/bstld","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"a-hierarchical-deep-architecture-and-mini","repo_url":"https://github.com/sovit-123/traffic-light-detection-using-yolov3","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":null,"task_name":"GPU"},{"task_slug":"traffic-sign-detection","task_name":"Traffic Sign Detection"},{"task_slug":"traffic-sign-recognition","task_name":"Traffic Sign Recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/traffic-sign-recognition-on-bosch-small","task":"Traffic Sign Recognition","dataset":"Bosch Small Traffic Lights","model":"Hierarchical + Background Threshold Model","rank_in_archive_order":1,"of":3,"metrics":{"MAP":"0.46"},"uses_additional_data":false},{"leaderboard":"/sota/traffic-sign-recognition-on-bosch-small","task":"Traffic Sign Recognition","dataset":"Bosch Small Traffic Lights","model":"Hierarchical Model","rank_in_archive_order":2,"of":3,"metrics":{"MAP":"0.45"},"uses_additional_data":false},{"leaderboard":"/sota/traffic-sign-recognition-on-bosch-small","task":"Traffic Sign Recognition","dataset":"Bosch Small Traffic Lights","model":"Background Threshold Model","rank_in_archive_order":3,"of":3,"metrics":{"MAP":"0.41"},"uses_additional_data":false},{"leaderboard":"/sota/traffic-sign-recognition-on-tsinghua-tencent","task":"Traffic Sign Recognition","dataset":"Tsinghua-Tencent 100K","model":"Background Threshold Model","rank_in_archive_order":3,"of":6,"metrics":{"MAP":"0.32"},"uses_additional_data":false},{"leaderboard":"/sota/traffic-sign-recognition-on-tsinghua-tencent","task":"Traffic Sign Recognition","dataset":"Tsinghua-Tencent 100K","model":"Hierarchical + Background Threshold Model","rank_in_archive_order":4,"of":6,"metrics":{"MAP":"0.31"},"uses_additional_data":false},{"leaderboard":"/sota/traffic-sign-recognition-on-tsinghua-tencent","task":"Traffic Sign Recognition","dataset":"Tsinghua-Tencent 100K","model":"Hierarchical Model","rank_in_archive_order":5,"of":6,"metrics":{"MAP":"0.30"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}