[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84974-en":3,"doc-seo-84974-105":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},84974,7971461740909,"Levi","https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d",8,"Research & Report","PUF Plug-and-Play Uncertainty-Aware Fusion for Online 3D Scene Graph Generation","Online 3D scene graph generation incrementally fuses partial 2D RGB-D observations into a persistent metric-space structure. Existing approaches assume deterministic fusion and ignore uncertainty from three areas: observation truncation, probabilistic ambiguity in 2D scene graph models, and approximate 3D lifting from noisy depth back-projection. PUF introduces a training-free, plug-and-play uncertainty-aware fusion framework that reformulates node association as probabilistic likelihood, accumulates Dirichlet evidence, optionally uses a class-conditional prior, and integrates both Gaussian and voxel backends.","arXiv :2607 .07 170v 1 [ cs .CV] 8 Jul 2026  \nPUF: Plug-and-Play Uncertainty-Aware Fusion for Online 3D Scene Graph Generation  \nYi Yang 1, Myrna Castillo2 ,3, Bodo Rosenhahn 1, and Michael Ying Yang3⋆  \n1 Leibniz Universität Hannover, Germany  \n2 Istituto Italiano di Tecnologia, Italy  \n3 University of Bath, UK  \nAbstract. Online 3D scene graph generation builds a persistent, structured representation of a scene by incrementally fusing 2D observations into a global 3D graph. Existing online methods treat this fusion asa fully deterministic pipeline, where we identify three sources of uncertainty that are overlooked: observation, 2D model, and 3D repre  \nsentation. We propose PUF: a Plug-and-play, Uncertainty-aware, and training-free Fusion framework. Scene graph node association is reformulated as a probabilistic likelihood over semantic and spatial factors, replacing binary accept/reject gates. Dirichlet evidence accumulation distributes class and relationship evidence across plausible candidates proportional to association likelihood. An optional class-conditional prior completes edges for sparsely or never co-observed object pairs. We instantiate PUF with both a 3D Gaussian and a 3D voxel backend and observe consistent improvements, demonstrating its ability to generalize across different representations. Experiments on the 3DSSG and ReplicaSSG benchmarks show that our method substantially outperforms existing approaches while maintaining real-time latency. These results establish uncertainty-aware fusion as a principled and effective paradigm for online 3D scene understanding. The source code is publicly available at [https://github.com/yyyyangyi/PUF](https://github.com/yyyyangyi/PUF).  \nKeywords: Scene graph generation · Uncertainty · Plug-and-Play  \n1 Introduction  \nA 2D Scene Graph (SG) describes an image as a structured representation in which nodes correspond to detected objects and directed edges encode their pairwise relationships. A 3D scene graph lifts this abstraction into metric space. The task of online 2D-to-3D Scene Graph Generation (SGG) is to construct a global 3D SG incrementally from a stream of RGB-D frames. 3D SGs provide the high-level scene abstraction needed for downstream tasks such as embodied navigation [28], robotic manipulation [45], and spatial question answering [39] . Incremental construction supports real-time applications for which a complete  \n⋆ Corresponding author.  \n2 Y. Yang et al.  \nscene reconstruction is unavailable or impractical [26], and lifting 2D predictions online into 3D leverages powerful 2D SGG models trained on richly annotated datasets [17] without the prohibitive cost of 3D SG annotation [33,34,42] .  \nIn the online setting, a global 3D SG is built by fusing a continuous stream of 2D observations, e.g ., a video sequence, into a persistent graph structure. With the camera’s field of view covering only a fraction of the scene at any instant, every such observation is partial and inherently uncertain. Accurately fusing these partial observations is non-trivial, as it requires reasoning about how much to trust each observation. However, existing online methods discard this information by committing to hard decisions at every stage. We identify three distinct sources of uncertainty, as illustrated in Fig. 1. (a) Observations are uncertain as objects are truncated. Relationships are particularly subject to such noise since they can only be predicted when both endpoint objects are simultaneously well observed. (b) 2D SGG models are uncertain. They produce soft class and relationship distributions that cannot be collapsed to hard predictions without loss.(c) Object representation in 3D is approximate. Back-projecting a 2D bounding box through a single noisy depth frame without committing to a full 3D reconstruction yields uncertain 3D position and spatial extent.  \nFig. 1: Our method is aware of 3 types of uncertainty which are overlooked in existing online 3D SGG methods. (a)","cbCaigJjXrQFx8I2","https://ap.wps.com/l/cbCaigJjXrQFx8I2","pdf",12189222,1,33,"English","en",105,"# Introduction\n## Online 2D-to-3D scene graph generation task\n## Sources of uncertainty in online fusion\n## Limitations of existing deterministic methods\n## PUF framework overview","[{\"question\":\"What is the core problem addressed by PUF in online 3D scene graph generation?\",\"answer\":\"PUF targets the challenge of fusing a stream of partial 2D observations into a global 3D scene graph while accounting for uncertainty instead of making hard deterministic decisions at each step.\"},{\"question\":\"Which three sources of uncertainty does the paper identify?\",\"answer\":\"The paper identifies uncertainty from (a) partial observations in 2D, especially truncation effects on objects and relationships, (b) uncertainty encoded by soft outputs from 2D SGG models, and (c) uncertainty introduced during 3D lifting from noisy depth back-projection and approximate 3D representations.\"},{\"question\":\"How does PUF incorporate uncertainty into the fusion process?\",\"answer\":\"PUF reformulates node association as a probabilistic likelihood over semantic and spatial factors, accumulates class and relationship evidence using Dirichlet evidence distribution, and optionally applies a class-conditional prior for sparsely or never co-observed object pairs.\"}]",1784199882,83,{"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},"puf-plug-and-play-uncertainty-aware-fusion-for-online-3d-scene-graph-generation","",{"@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/puf-plug-and-play-uncertainty-aware-fusion-for-online-3d-scene-graph-generation/84974/",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","2026-07-17","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},"What is the core problem addressed by PUF in online 3D scene graph generation?","Question",{"text":75,"@type":76},"PUF targets the challenge of fusing a stream of partial 2D observations into a global 3D scene graph while accounting for uncertainty instead of making hard deterministic decisions at each step.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Which three sources of uncertainty does the paper identify?",{"text":80,"@type":76},"The paper identifies uncertainty from (a) partial observations in 2D, especially truncation effects on objects and relationships, (b) uncertainty encoded by soft outputs from 2D SGG models, and (c) uncertainty introduced during 3D lifting from noisy depth back-projection and approximate 3D representations.",{"name":82,"@type":73,"acceptedAnswer":83},"How does PUF incorporate uncertainty into the fusion process?",{"text":84,"@type":76},"PUF reformulates node association as a probabilistic likelihood over semantic and spatial factors, accumulates class and relationship evidence using Dirichlet evidence distribution, and optionally applies a class-conditional prior for sparsely or never co-observed object pairs.","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"]