[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85360-en":3,"doc-seo-85360-105":29,"detail-sidebar-cat-0-en-105":91},{"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":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},85360,7971461741311,"Ophelia","https://ap-avatar.wpscdn.com/avatar/74000253aff267980c6?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779345379180704826",8,"Research & Report","Higher-Order Cell Tracking Transformer","Reconstructing cell lineages from live-imaging microscopy requires linking detections across time, including through divisions. Many candidate-graph tracking methods miss key structural obstacles: division-induced entanglement in node embeddings and near-random label agreement among edges incident to the same node, leaving graph topology unhelpful for aggregation. The Higher-Order Cell Tracking Transformer (HOCT) uses an edge-centric design where candidate links attend via a 3D geometric prior, addressing both issues. Evaluations on the Cell Tracking Challenge and a bacteria division benchmark show state-of-the-art accuracy without deep pre-trained image encoders, and it fine-tunes faster with substantial error reduction in human-in-the-loop settings.","arXiv :2607 . 11754v1 [ cs .CV] 13 Jul 2026  \nHigher-Order Cell Tracking Transformer  \nJordão Bragantini Ilan Theodoro Loïc A. Royer  \nBiohub  \nSan Francisco, CA  \n{jordao.bragantini, ilan.silva, [loic.royer}@biohub.org](loic.royer}@biohub.org)  \nAbstract  \nReconstructing lineages from live-imaging microscopy requires linking cell detections across time, including through cell divisions. A common approach is to construct a candidate graph and associate cell segmentations (nodes) across frames. However, these and other existing methods overlook two structural obstacles in candidate tracking graphs: (i) cell divisions entangle distinct lineage pathsin the node embedding space, and (ii) edges sharing a node have near-random label agreement, so the candidate-graph topology carries no useful information for graph neural networks to aggregate. We propose the Higher-Order Cell Tracking Transformer (HOCT), an edge-centric architecture in which candidate cell links attend to one another under a 3D geometric prior, resolving both issues. Evaluated on the Cell Tracking Challenge and a bacteria division benchmark, HOCT achieves state-of-the-art results without deep pre-trained image encoders. Moreover, the proposed approach is easier to fine-tune, quickly reducing tracking errors by 59% with  \n400 annotations in a human-in-the-loop setting, outperforming LoRA fine-tuning of competing transformer baselines (6.75% improvement) .  \n1 Introduction  \nSince the earliest microscopes [21], advances in imaging have been inseparable from advances in biology. Modern live-imaging platforms [26, 12, 6, 56, 47] routinely produce terabytes of volumetric time-lapse data, enabling high-content screening [15], the study of embryonic development [31], tissue regeneration [11], and more. A fundamental computational challenge in these applications is cell tracking: the reconstruction of cell lineages, capturing trajectories and division events over time. The current dominant paradigm is tracking by detection [23, 36], in which cells are first segmented in each frame and subsequently linked into lineages. Unlike multi-object tracking in natural images [13], cells are visually near-identical and lack persistent appearance cues, impairing re-identification methods [55, 39] that succeed in pedestrian or vehicle tracking [4, 3, 38] . Worse, cells can divide, changing their morphology and producing a variable number of objects. Cell tracking has therefore relied on specialized solutions that exploit spatial information, either explicitly via optical-flow [19, 35, 44] or implicitly by feeding appearance and positional features to learned models [2, 16] . Recent learning-based approaches employ neural networks on spatiotemporal graphs. Trackastra [16] uses an encoder-decoder Transformer with multi-dimensional Rotary Position Embeddings (RoPE) [48, 20] to embed each cell detection (node) into a latent space and compute association probabilities from cosine similarity between all nodes in adjacent frames. Ben-Haim and Raviv [2] and Braso and Leal-Taixe [9] formulate tracking as edge classification on spatiotemporal graphs, for cell and pedestrian tracking respectively, [2] using GNN message passing on the candidate graph [45] . Node-embedding approaches face two structural problems on this graph (Fig. 1a–b) . First, divisions create a connected-manifold effect: when a cell p divides into daughters d 1 , d2 , the embedding of p must be close to both, merging lineage paths that should form separate clusters and leaving them  \nPreprint.  \nFigure 1: Motivation and architecture. (a) A candidate tracking graph and its ground-truth solution, where each color denotes a distinct cell trajectory (simple path in the lineage tree) . The graph is non-homophilic: edges sharing a node have near-random label agreement, so graph topology carries no useful label information. (b) In node embedding approaches, divisions create connected manifolds that merge distinct lineage paths, confoun","cbCaiguq8RVJyPpV","https://ap.wps.com/l/cbCaiguq8RVJyPpV","pdf",992492,1,26,"English","en",105,"# Abstract\n# Introduction\n## Cell tracking background and challenges\n## Limitations of node-embedding and GNN candidate graphs\n## HOCT edge-centric architecture and contributions","[{\"question\":\"Why do conventional candidate graph tracking methods struggle with cell divisions?\",\"answer\":\"Divisions create embedding entanglement, merging lineage paths in node embedding space and confounding similarity-based association. HOCT resolves this by operating in an edge-centric space with geometry-biased attention.\"},{\"question\":\"What does “non-homophilic” mean in candidate tracking graphs, and how does it affect learning?\",\"answer\":\"Only a small fraction of edges incident to a node are correct, so edges sharing a node have near-random label agreement. This causes message passing to aggregate noise, limiting GNN performance.\"},{\"question\":\"What is the core idea of HOCT and how does it differ from node-embedding approaches?\",\"answer\":\"HOCT treats each candidate cell link as an independent token and refines edge representations through attention biased by inter-edge geometry, rather than relying on candidate-graph adjacency or node embedding similarity.\"}]",1784202775,66,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"higher-order-cell-tracking-transformer","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/higher-order-cell-tracking-transformer/85360/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-22","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why do conventional candidate graph tracking methods struggle with cell divisions?","Question",{"text":75,"@type":76},"Divisions create embedding entanglement, merging lineage paths in node embedding space and confounding similarity-based association. HOCT resolves this by operating in an edge-centric space with geometry-biased attention.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What does “non-homophilic” mean in candidate tracking graphs, and how does it affect learning?",{"text":80,"@type":76},"Only a small fraction of edges incident to a node are correct, so edges sharing a node have near-random label agreement. This causes message passing to aggregate noise, limiting GNN performance.",{"name":82,"@type":73,"acceptedAnswer":83},"What is the core idea of HOCT and how does it differ from node-embedding approaches?",{"text":84,"@type":76},"HOCT treats each candidate cell link as an independent token and refines edge representations through attention biased by inter-edge geometry, rather than relying on candidate-graph adjacency or node embedding similarity.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]