{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,5,13]],"date-time":"2025-05-13T13:59:25Z","timestamp":1747144765538,"version":"3.40.5"},"reference-count":31,"publisher":"Wiley","issue":"4","license":[{"start":{"date-parts":[[2023,11,21]],"date-time":"2023-11-21T00:00:00Z","timestamp":1700524800000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["AI Magazine"],"published-print":{"date-parts":[[2023,12]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:p>As autonomous vehicle technology advances, high\u2010definition (HD) maps have become essential for ensuring safety and navigation accuracy. However, creating HD maps with accurate annotations demands substantial human effort, leading to a time\u2010consuming and costly process. Although artificial intelligence (AI) and computer vision (CV) algorithms have been developed for prelabeling HD maps, a significant gap remains in accuracy and robustness between AI\u2010based methods and traditional manual pipelines. Additionally, building large\u2010scale annotated datasets and advanced machine learning algorithms for AI\u2010based HD map labeling systems can be resource\u2010intensive. In this paper, we present and summarize the Tencent HD Map AI (THMA) system, an innovative end\u2010to\u2010end, AI\u2010based, active learning HD map labeling system designed to produce HD map labels for hundreds of thousands of kilometers while employing active learning to enhance product iteration. Utilizing a combination of supervised, self\u2010supervised, and weakly supervised learning, THMA is trained directly on massive HD map datasets to achieve the high accuracy and efficiency required by downstream users. Deployed by the Tencent Map team, THMA serves over 1000 labeling workers and generates more than 30,000 km of HD map data per day at its peak. With over 90% of Tencent Map's HD map data labeled automatically by THMA, the system accelerates traditional HD map labeling processes by more than tenfold, significantly reducing manual annotation burdens and paving the way for more efficient HD map\u00a0production.<\/jats:p>","DOI":"10.1002\/aaai.12139","type":"journal-article","created":{"date-parts":[[2023,11,22]],"date-time":"2023-11-22T19:42:21Z","timestamp":1700682141000},"page":"418-430","update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["High\u2010definition map automatic annotation system based on active learning"],"prefix":"10.1002","volume":"44","author":[{"given":"Chao","family":"Zheng","sequence":"first","affiliation":[{"name":"T Lab Tencent  Beijing China"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8739-5196","authenticated-orcid":false,"given":"Xu","family":"Cao","sequence":"additional","affiliation":[{"name":"University of Illinois at Urbana\u2010Champaign  Champaign Illinois USA"}]},{"given":"Kun","family":"Tang","sequence":"additional","affiliation":[{"name":"T Lab Tencent  Beijing China"}]},{"given":"Zhipeng","family":"Cao","sequence":"additional","affiliation":[{"name":"T Lab Tencent  Beijing China"}]},{"given":"Elena","family":"Sizikova","sequence":"additional","affiliation":[{"name":"New York University  New York New York USA"}]},{"given":"Tong","family":"Zhou","sequence":"additional","affiliation":[{"name":"T Lab Tencent  Beijing China"}]},{"given":"Erlong","family":"Li","sequence":"additional","affiliation":[{"name":"T Lab Tencent  Beijing China"}]},{"given":"Ao","family":"Liu","sequence":"additional","affiliation":[{"name":"T Lab Tencent  Beijing China"}]},{"given":"Shengtao","family":"Zou","sequence":"additional","affiliation":[{"name":"T Lab Tencent  Beijing China"}]},{"given":"Xinrui","family":"Yan","sequence":"additional","affiliation":[{"name":"T Lab Tencent  Beijing China"}]},{"given":"Shuqi","family":"Mei","sequence":"additional","affiliation":[{"name":"T Lab Tencent  Beijing China"}]}],"member":"311","published-online":{"date-parts":[[2023,11,21]]},"reference":[{"key":"e_1_2_7_2_1","unstructured":"Bao Hangbo LiDong SonghaoPiao andFuruWei.2021. \u201cBEiT: BERT Pre\u2010Training of Image Transformers.\u201d InInternational Conference on Learning Representations."},{"key":"e_1_2_7_3_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.engappai.2023.106125"},{"key":"e_1_2_7_4_1","doi-asserted-by":"crossref","unstructured":"Caesar Holger VarunBankiti AlexH Lang SourabhVora VeniceErin Liong QiangXu AnushKrishnan YuPan GiancarloBaldan andOscarBeijbom.2020. \u201cnuScenes: A Multimodal Dataset for Autonomous Driving.\u201d InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition 11621\u201311631.","DOI":"10.1109\/CVPR42600.2020.01164"},{"key":"e_1_2_7_5_1","unstructured":"Dosovitskiy Alexey LucasBeyer AlexanderKolesnikov DirkWeissenborn XiaohuaZhai ThomasUnterthiner MostafaDehghani etal.2020. \u201cAn Image is Worth 16x16 Words: Transformers for Image Recognition at Scale.\u201d InInternational Conference on Learning Representations."},{"key":"e_1_2_7_6_1","doi-asserted-by":"crossref","unstructured":"Elhousni Mahdi YechengLyu ZimingZhang andXinmingHuang.2020. \u201cAutomatic Building and Labeling of HD Maps with Deep Learning.\u201d InProceedings of the AAAI Conference on Artificial Intelligence volume34 13255\u201313260.","DOI":"10.1609\/aaai.v34i08.7033"},{"key":"e_1_2_7_7_1","unstructured":"Fan Haoyang FanZhu ChangchunLiu LiangliangZhang LiZhuang DongLi WeichengZhu JiangtaoHu HongyeLi andQiKong.2018. \u201cBaidu Apollo Em Motion Planner.\u201darXiv preprint arXiv:1807.08048.https:\/\/arxiv.org\/abs\/1807.08048 https:\/\/github.com\/ApolloAuto\/apollo"},{"key":"e_1_2_7_8_1","first-page":"35946","article-title":"Masked autoencoders as spatiotemporal learners","volume":"35","author":"Feichtenhofer Christoph","year":"2022","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_7_9_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.inffus.2020.11.002"},{"key":"e_1_2_7_10_1","doi-asserted-by":"crossref","unstructured":"Haussmann Elmar MicheleFenzi KashyapChitta JanIvanecky HansonXu DonnaRoy AkshitaMittel NicolasKoumchatzky ClementFarabet andJoseM Alvarez.2020. \u201cScalable Active Learning for Object Detection.\u201d In2020 IEEE Intelligent Vehicles Symposium (IV) 1430\u20131435.IEEE.","DOI":"10.1109\/IV47402.2020.9304793"},{"key":"e_1_2_7_11_1","doi-asserted-by":"crossref","unstructured":"He Kaiming XinleiChen SainingXie YanghaoLi PiotrDoll\u00e1r andRossGirshick.2022. \u201cMasked Autoencoders are Scalable Vision Learners.\u201d InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition 16000\u201316009.","DOI":"10.1109\/CVPR52688.2022.01553"},{"key":"e_1_2_7_12_1","doi-asserted-by":"crossref","unstructured":"Hou Yuenan XingeZhu YuexinMa ChenChange Loy andYikangLi.2022. \u201cPoint\u2010To\u2010Voxel Knowledge Distillation for Lidar Semantic Segmentation.\u201d InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition 8479\u20138488.","DOI":"10.1109\/CVPR52688.2022.00829"},{"key":"e_1_2_7_13_1","doi-asserted-by":"crossref","unstructured":"Jiao Jialin.2018. \u201cMachine Learning Assisted High\u2010Definition Map Creation.\u201d In2018 IEEE 42nd Annual Computer Software and Applications Conference (COMPSAC) volume1 367\u2013373.IEEE.","DOI":"10.1109\/COMPSAC.2018.00058"},{"key":"e_1_2_7_14_1","doi-asserted-by":"publisher","DOI":"10.1109\/LRA.2021.3060406"},{"key":"e_1_2_7_15_1","doi-asserted-by":"crossref","unstructured":"Li Qi YueWang YilunWang andHangZhao.2022. \u201cHDMapNet: An Online HD Map Construction and Evaluation Framework.\u201d In2022 International Conference on Robotics and Automation (ICRA) 4628\u20134634.IEEE.","DOI":"10.1109\/ICRA46639.2022.9812383"},{"key":"e_1_2_7_16_1","doi-asserted-by":"crossref","unstructured":"Long Jonathan EvanShelhamer andTrevorDarrell.2015. \u201cFully Convolutional Networks for Semantic Segmentation.\u201d InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition 3431\u20133440.","DOI":"10.1109\/CVPR.2015.7298965"},{"key":"e_1_2_7_17_1","doi-asserted-by":"crossref","unstructured":"M\u00e1ttyus Gell\u00e9rt ShenlongWang SanjaFidler andRaquelUrtasun.2016. \u201cHD Maps: Fine\u2010Grained Road Segmentation by Parsing Ground and Aerial Images.\u201d InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition 3611\u20133619.","DOI":"10.1109\/CVPR.2016.393"},{"key":"e_1_2_7_18_1","doi-asserted-by":"crossref","unstructured":"Pannen David MartinLiebner WolfgangHempel andWolframBurgard.2020. \u201cHow to Keep HD Maps for Automated Driving Up to Date.\u201d In2020 IEEE International Conference on Robotics and Automation (ICRA) 2288\u20132294.IEEE.","DOI":"10.1109\/ICRA40945.2020.9197419"},{"key":"e_1_2_7_19_1","doi-asserted-by":"crossref","unstructured":"Ronneberger Olaf PhilippFischer andThomasBrox.2015. \u201cU\u2010Net: Convolutional Networks for Biomedical Image Segmentation.\u201d InInternational Conference on Medical Image Computing and Computer\u2010assisted Intervention 234\u2013241.Springer.","DOI":"10.1007\/978-3-319-24574-4_28"},{"key":"e_1_2_7_20_1","doi-asserted-by":"crossref","unstructured":"Sun Pei HenrikKretzschmar XerxesDotiwalla AurelienChouard VijaysaiPatnaik PaulTsui JamesGuo etal.2020. \u201cScalability in Perception for Autonomous Driving: Waymo Open Dataset.\u201d InProceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition 2446\u20132454.","DOI":"10.1109\/CVPR42600.2020.00252"},{"key":"e_1_2_7_21_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2020.107623"},{"key":"e_1_2_7_22_1","doi-asserted-by":"crossref","unstructured":"Tang Kun XuCao ZhipengCao TongZhou ErlongLi AoLiu ShengtaoZou etal.2023. \u201cTHMA: Tencent HD Map Ai System for Creating HD Map Annotations.\u201d InProceedings of the AAAI Conference on Artificial Intelligence volume37 15585\u201315593.","DOI":"10.1609\/aaai.v37i13.26848"},{"key":"e_1_2_7_23_1","unstructured":"Tao Andrew KaranSapra andBryanCatanzaro.2020. \u201cHierarchical Multi\u2010Scale Attention for Semantic Segmentation.\u201darXiv preprint arXiv:2005.10821.https:\/\/arxiv.org\/abs\/2005.10821"},{"key":"e_1_2_7_24_1","first-page":"10078","article-title":"Videomae: Masked autoencoders are data\u2010efficient learners for self\u2010supervised video pre\u2010training","volume":"35","author":"Tong Zhan","year":"2022","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_7_25_1","doi-asserted-by":"crossref","unstructured":"Wang Wenhai EnzeXie XiangLi Deng\u2010PingFan KaitaoSong DingLiang TongLu PingLuo andLingShao.2021. \u201cPyramid Vision Transformer: A Versatile Backbone for Dense Prediction Without Convolutions.\u201d InProceedings of the IEEE\/CVF International Conference on Computer Vision 568\u2013578.","DOI":"10.1109\/ICCV48922.2021.00061"},{"key":"e_1_2_7_26_1","unstructured":"Wilson Benjamin WilliamQi TanmayAgarwal JohnLambert JagjeetSingh SiddheshKhandelwal BowenPan etal.2021. \u201cArgoverse 2: Next Generation Datasets for Self\u2010Driving Perception and Forecasting.\u201d InThirty\u2010fifth Conference on Neural Information Processing Systems Datasets and Benchmarks Track (Round 2)."},{"key":"e_1_2_7_27_1","first-page":"12077","article-title":"SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers","volume":"34","author":"Xie Enze","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_7_28_1","doi-asserted-by":"crossref","unstructured":"Yan Chao ChaoZheng ChaoGao WeiYu YuzhanCai andChangjieMa.2020. \u201cLane Information Perception Network for HD Maps.\u201d In2020 IEEE 23rd International Conference on Intelligent Transportation Systems (ITSC) 1\u20136.IEEE.","DOI":"10.1109\/ITSC45102.2020.9294666"},{"key":"e_1_2_7_29_1","unstructured":"Yang Bin MingLiang andRaquelUrtasun.2018. \u201cHDNET: Exploiting HD Maps for 3D Object Detection.\u201d InConference on Robot Learning 146\u2013155.PMLR."},{"key":"e_1_2_7_30_1","doi-asserted-by":"crossref","unstructured":"Yang Bin WenjieLuo andRaquelUrtasun.2018. \u201cPixor: Real\u2010Time 3D Object Detection from Point Clouds.\u201d InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition 7652\u20137660.","DOI":"10.1109\/CVPR.2018.00798"},{"key":"e_1_2_7_31_1","doi-asserted-by":"crossref","unstructured":"Zhou Yiyang YuichiTakeda MasayoshiTomizuka andWeiZhan.2021. \u201cAutomatic Construction of Lane\u2010Level HD Maps for Urban Scenes.\u201d In2021 IEEE\/RSJ International Conference on Intelligent Robots and Systems (IROS) 6649\u20136656.IEEE.","DOI":"10.1109\/IROS51168.2021.9636205"},{"key":"e_1_2_7_32_1","doi-asserted-by":"publisher","DOI":"10.1093\/nsr\/nwx106"}],"container-title":["AI Magazine"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/aaai.12139","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,12,24]],"date-time":"2023-12-24T05:42:42Z","timestamp":1703396562000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1002\/aaai.12139"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,11,21]]},"references-count":31,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2023,12]]}},"alternative-id":["10.1002\/aaai.12139"],"URL":"https:\/\/doi.org\/10.1002\/aaai.12139","archive":["Portico"],"relation":{},"ISSN":["0738-4602","2371-9621"],"issn-type":[{"type":"print","value":"0738-4602"},{"type":"electronic","value":"2371-9621"}],"subject":[],"published":{"date-parts":[[2023,11,21]]},"assertion":[{"value":"2023-04-24","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-06-27","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-11-21","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}