[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85238-en":3,"doc-seo-85238-105":30,"detail-sidebar-cat-0-en-105":92},{"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":13,"seo_description":14,"update_tm":28,"read_time":29},85238,1374391974564,"Clementine","https://ap-avatar.wpscdn.com/avatar/14000253aa45c000a9e?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779874745381141002",8,"Research & Report","Incremental Online Scene Reconstruction by 3D Gaussian Triangulation","Incremental scene reconstruction is essential for real-world applications, yet most 3D Gaussian Splatting approaches require offline conversion of optimized Gaussians into an intermediate implicit field for explicit mesh extraction, limiting seamless downstream integration. The work proposes an online framework that incrementally reconstructs and updates high-fidelity explicit meshes by directly triangulating a dense Gaussian representation. A direct meshing algorithm extracts and updates meshes efficiently, while a plane-based pulling constraint aligns Gaussian primitives to the local surface and reduces memory/computation via dynamic freezing of optimized historical regions. Experiments on public datasets show improved rendering quality and reconstruction accuracy.","arXiv :2607 . 10690v1 [ cs .CV] 12 Jul 2026  \nIncremental Online Scene Reconstruction by 3D Gaussian Triangulation  \nYanjin Zhu* 1, Shaofan Liu* 2, and Jianke Zhu†1 ,3  \n1 Zhejiang University, Hangzhou, China  \n2 Hefei University of Technology, Hefei, China  \n3 Shenzhen Loop Area Institute, Shenzhen, China  \nAbstract. Incremental scene reconstruction is essential for real-world applications. Although 3D Gaussian Splatting shows strong potential, most existing approaches require offline conversion of the optimized Gaussians into an intermediate implicit field for explicit mesh extraction, which hinders seamless integration with downstream tasks. To address this limitation, we propose a novel online framework that incrementallyreconstructs and updates high-fidelity explicit meshes by directly triangulating a dense geometric Gaussian representation, which supports both high-quality rendering and incremental surface reconstruction. Moreover, we present a direct meshing algorithm that efficiently extractsand updates the mesh from the Gaussian set. To ensure mesh accuracy, we enforce a plane-based pulling constraint that dynamically aligns 3D Gaussian primitives to the approximated local surface. Furthermore, our framework significantly reduces memory and computational overhead during long-sequence processing by dynamically freezing fully optimized historical regions. Experiments on public datasets demonstrate that our method outperforms conventional Gaussian-based methods on both rendering quality and reconstruction accuracy.  \nKeywords: Gaussian Triangulation · Incremental Reconstruction · Gaussian Splatting  \n1 Introduction  \n3D scene reconstruction is a fundamental task with extensive applications in computer vision and robotics, notably Augmented Reality [21] and scene perception [8] . Typically, 3D scene reconstruction techniques can be categorized into two groups. One is implicit methods [17, 28], and the other is explicit approaches [7,12,27] . Unfortunately, most high-fidelity reconstruction methods require processing the entire scene before generating the recovered meshes, which poses a significant bottleneck in real-world applications. For example, the memory constraints of computing devices make it infeasible to represent continuously expanding scenes. For time-critical tasks in autonomous robotics, the delay in  \n*  \n†  \nEqual contribution. Corresponding author.  \n2 Y. Zhu et al.  \nGlobal Optimization  \nMesh Extraction  \nFig. 1: Comparison between traditional batch reconstruction and our proposed incremental framework. Left: Traditional methods typically rely on global optimization overall available frames (T1...TN ) . Once a new frame (TN+1) is obtained, the entire mesh extraction pipeline has to be recomputed from scratch. Right: Our approach performsin a fully incremental manner. As the input stream arrives (T1 ... TN+M), we progressively update the Gaussian representation and the reconstructed mesh. This allows for continuous scene expansion without global re-optimization.  \nobtaining a complete model is impractical, as early access to partial results is crucial for making immediate decisions. Consequently, incremental 3D scene reconstruction is urgently needed to overcome these fundamental limitations.  \nTraditional online scene reconstruction methods, such as KinectFusion [18] and its derivatives, are typically based on volumetric integration of the Truncated Signed Distance Field (TSDF) . These approaches incrementally fuse sequential depth maps to update a volumetric grid, from which a triangle mesh is subsequently extracted by Marching Cubes [13] at a much lower frequency than the TSDF fusion. However, they are often constrained by fixed voxel resolution and high memory consumption, which prevent detailed capture and large-scale scene reconstruction.  \nA remedy is to utilize continuous neural implicit representations, notably Neural Radiance Fields (NeRF) [17], which models scenes as radiance fields with adapti","cbCaihQoe8H2FJv8","https://ap.wps.com/l/cbCaihQoe8H2FJv8","pdf",3635584,5,1,17,"English","en",105,"# Introduction\n## Background and Motivation\n## Limitations of Existing Online and Implicit Methods\n## Related Work on 3D Gaussian Splatting and Mesh Extraction\n## Proposed Online Incremental Framework","[{\"question\":\"What limitation does the proposed method address in existing 3D Gaussian Splatting approaches?\",\"answer\":\"Most existing methods require offline conversion of optimized Gaussians into an intermediate implicit field to enable explicit mesh extraction, which blocks seamless integration with downstream tasks.\"},{\"question\":\"How does the framework perform incremental reconstruction online?\",\"answer\":\"It treats optimized 3D Gaussians as surfel primitives and directly triangulates a dense Gaussian representation, incrementally updating both the Gaussian model and the reconstructed mesh as new frames arrive.\"},{\"question\":\"What mechanism is used to improve mesh accuracy from the Gaussian primitives?\",\"answer\":\"A plane-based pulling constraint dynamically aligns 3D Gaussian primitives to the approximated local surface, ensuring the extracted mesh remains accurate.\"}]",1784201940,43,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":28},"incremental-online-scene-reconstruction-by-3d-gaussian-triangulation","",{"@graph":36,"@context":86},[37,54,69],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/incremental-online-scene-reconstruction-by-3d-gaussian-triangulation/85238/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-07-24","2026-07-16",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"What limitation does the proposed method address in existing 3D Gaussian Splatting approaches?","Question",{"text":76,"@type":77},"Most existing methods require offline conversion of optimized Gaussians into an intermediate implicit field to enable explicit mesh extraction, which blocks seamless integration with downstream tasks.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does the framework perform incremental reconstruction online?",{"text":81,"@type":77},"It treats optimized 3D Gaussians as surfel primitives and directly triangulates a dense Gaussian representation, incrementally updating both the Gaussian model and the reconstructed mesh as new frames arrive.",{"name":83,"@type":74,"acceptedAnswer":84},"What mechanism is used to improve mesh accuracy from the Gaussian primitives?",{"text":85,"@type":77},"A plane-based pulling constraint dynamically aligns 3D Gaussian primitives to the approximated local surface, ensuring the extracted mesh remains accurate.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,110,115,120,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":20,"slug":138},19,"General","general"]