[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-125872-en":3,"doc-seo-125872-105":31,"detail-sidebar-cat-0-en-105":93},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":28,"seo_description":14,"update_tm":29,"read_time":30},125872,1099523885336,"Violet","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","BESIII轨迹重建算法：基于机器学习","Track reconstruction is a central yet demanding task in offline collider data processing, especially for the BESIII detector operating in the tau-charm energy region. While traditional approaches like template matching and Hough transform have improved tracking performance, challenging cases such as low-momentum tracks, secondary-vertex tracks, and high-noise environments still leave significant room for progress. This work presents a machine-learning-based tracking pipeline using hit-pattern maps, a graph neural network for hit-on-track versus noise classification, DBSCAN combined with RANSAC for multi-track clustering, and Genfit2 for track fitting with deterministic annealing filtering.","BESIII track reconstruction algorithm based on machine learning  \nXiaoqian Jia1 , Xiaoshuai Qin1 ∗ , Teng Li1 ∗∗ , Xingtao Huang1 ∗∗∗ , Xueyao Zhang1 , Na Yin 1 , Yao Zhang2 , and Ye Yuan2  \n1 Shandong University, 226237, Qingdao, China  \n2Institute of High Energy Physics, Chinese Academy of Sciences, 100049, Beijing, China  \nAbstract. Track reconstruction is one of the most important and challenging tasks in the offline data processing of collider experiments. For the BESIII detector working in the tau-charm energy region, plenty of efforts were made previously to improve the tracking performance with traditional methods, such as template matching and Hough transform etc. However, for difficult tracking tasks, such as the tracking of low momentum tracks, tracks from secondary vertices and tracks with high noise level, there is still large room for improvement.  \nIn this contribution, we demonstrate a novel tracking algorithm based on machine learning method. In this method, a hit pattern map representing the connectivity between drift cells is established using an enormous MC sample, based on which we design an optimal method of graph construction, then an edgeclassifying Graph Neural Network is trained to distinguish the hit-on-track from noise hits. Finally, a clustering method based on DBSCAN and RANSAC is developed to cluster hits from multiple tracks. Track fitting algorithm based on GENFIT2 is also studied to obtain the track parameters, where deterministic annealing filter are implemented to deal with ambiguities and potential noises.  \nThe preliminary results on BESIII MC sample presents promising performance, showing potential to apply this method to other trackers based on drift chamber as well, such as the CEPC and STCF detectors under pre-study.  \n1 Introduction  \nBeijing Spectrometer III (BESIII) [1] is a magnetic spectrometer detector operating at BEPCII with √s = 2 ∼ 4.9 GeV/c2. BESIII’s primary physics goals are to study theeletroweak and strong interactions and to search for new physics at the tau-charm energy region. At the boundary between the perturbative and nonperturbative regimes of QCD, BESIII offers vast and diverse physics opportunities.Plenty results related to hadron physics and τ-charm physics play an important role in the understanding of particle’s inner structure and interactions [2] .  \nThe Main Drift Chamber (MDC) is a crucial component of the BESIII detector, which is mainly responsible for the precise detection of the trajectory of charged particles. MDC is a cylindrical detector consisting of 43 layers of drift cells filled with a gas mixture. When  \n∗ e-mail: [qinxs@ihep.ac.cn](qinxs@ihep.ac.cn)  \n∗∗ e-mail: [tengli@sdu.edu.cn](tengli@sdu.edu.cn)  \n∗∗∗ e-mail: [huangxt@sdu.edu.cn](huangxt@sdu.edu.cn)  \n© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 ([https://creativecommons.org/licenses/by/4.0/](https://creativecommons.org/licenses/by/4.0/)).  \na charged particle passes through the MDC, it ionizes the gas atoms, creating electron-ion pairs. The electric field within the drift cells causes the electrons to drift towards the anode wires, while the ions drift towards the cathode plates. By measuring the arrival times of the electrons at the anode wires, the MDC is able to detect the hits of the particle along its trajectory.  \nOver the last decades, traditional tracking strategy for the BESIII MDC system has been successfully working as the official reconstruction algorithms, including the tracking finding algorithms via PATTSF [3, 4] and HOUGH [5], track parameters estimation based on Kalman Fitter and Rugge Kutta methods. This strategy works well, but there is still a decent room for improvement, especially for tracks with low momentum, high noise level as well as tracks from long lived particles (secondary vertex) .  \nInspired by the work on TrackML particle tracking challenge, which is a","cbCaiqkLN4jniyn7","https://ap.wps.com/l/cbCaiqkLN4jniyn7","pdf",2192859,5,1,7,"English","en",105,"# Introduction\n## Filtering noise via GNN\n### Graph construction\n# (Further sections)\n## Track finding with DBSCAN and RANSAC\n## Track fitting and kinematic parameter extraction\n## Summary and outlook","[{\"question\":\"Why is track reconstruction difficult for BESIII in the tau-charm energy region?\",\"answer\":\"Because challenging scenarios such as low-momentum tracks, secondary-vertex tracks, and high-noise levels make reliable hit-to-track association harder, even after improvements from traditional methods.\"},{\"question\":\"How does the proposed algorithm classify noise hits?\",\"answer\":\"It builds a hit pattern map for graph construction from a large MC sample, then trains an edge-classifying graph neural network to distinguish hit-on-track from noise hits.\"},{\"question\":\"How are hits clustered into tracks and how are track parameters obtained?\",\"answer\":\"Hits from multiple tracks are clustered using DBSCAN together with RANSAC, and then track parameters are extracted using Genfit2 with deterministic annealing filtering to address ambiguities and potential noises.\"}]","BESIII轨迹重建算法：基于机器学习 | PDF",1785901740,18,{"code":4,"msg":32,"data":33},"ok",{"site_id":25,"language":24,"slug":34,"title":13,"keywords":35,"description":14,"schema_data":36,"social_meta":88,"head_meta":90,"extra_data":92,"updated_unix":29},"besiii-track-reconstruction-algorithm-based-on-machine-learning","",{"@graph":37,"@context":87},[38,55,70],{"@type":39,"itemListElement":40},"BreadcrumbList",[41,45,49,52],{"item":42,"name":43,"@type":44,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":46,"name":47,"@type":44,"position":48},"https://docshare.wps.com/document/","Document",2,{"item":50,"name":12,"@type":44,"position":51},"https://docshare.wps.com/document/research-report/",3,{"item":53,"name":13,"@type":44,"position":54},"https://docshare.wps.com/document/besiii-track-reconstruction-algorithm-based-on-machine-learning/125872/",4,{"url":53,"name":13,"@type":56,"author":57,"headline":13,"publisher":59,"fileFormat":62,"inLanguage":24,"description":14,"dateModified":63,"datePublished":64,"encodingFormat":62,"isAccessibleForFree":65,"interactionStatistic":66},"DigitalDocument",{"name":9,"@type":58},"Person",{"url":42,"name":60,"@type":61},"DocShare","Organization","application/pdf","2026-08-24","2026-08-05",true,{"@type":67,"interactionType":68,"userInteractionCount":20},"InteractionCounter",{"@type":69},"ViewAction",{"@type":71,"mainEntity":72},"FAQPage",[73,79,83],{"name":74,"@type":75,"acceptedAnswer":76},"Why is track reconstruction difficult for BESIII in the tau-charm energy region?","Question",{"text":77,"@type":78},"Because challenging scenarios such as low-momentum tracks, secondary-vertex tracks, and high-noise levels make reliable hit-to-track association harder, even after improvements from traditional methods.","Answer",{"name":80,"@type":75,"acceptedAnswer":81},"How does the proposed algorithm classify noise hits?",{"text":82,"@type":78},"It builds a hit pattern map for graph construction from a large MC sample, then trains an edge-classifying graph neural network to distinguish hit-on-track from noise hits.",{"name":84,"@type":75,"acceptedAnswer":85},"How are hits clustered into tracks and how are track parameters obtained?",{"text":86,"@type":78},"Hits from multiple tracks are clustered using DBSCAN together with RANSAC, and then track parameters are extracted using Genfit2 with deterministic annealing filtering to address ambiguities and potential noises.","https://schema.org",{"og:url":53,"og:type":89,"og:title":13,"og:site_name":60,"og:description":14},"article",{"robots":91,"canonical":53},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":94},[95,99,103,107,111,116,120,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":47,"category_name":96,"show_sort_weight":97,"slug":98},"Story & Novel",90,"story-novel",{"id":48,"doc_module":4,"doc_module_name":47,"category_name":100,"show_sort_weight":101,"slug":102},"Literature",80,"literature",{"id":54,"doc_module":4,"doc_module_name":47,"category_name":104,"show_sort_weight":105,"slug":106},"Exam",70,"exam",{"id":20,"doc_module":4,"doc_module_name":47,"category_name":108,"show_sort_weight":109,"slug":110},"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":47,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":22,"doc_module":4,"doc_module_name":47,"category_name":117,"show_sort_weight":118,"slug":119},"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":47,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":47,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":47,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":47,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":47,"category_name":137,"show_sort_weight":20,"slug":138},19,"General","general"]