[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82273-en":3,"doc-seo-82273-105":30,"detail-sidebar-cat-0-en-105":90},{"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},82273,962075114101,"Seraphina","https://ap-avatar.wpscdn.com/avatar/e000253a75eb197efd?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780044092746381165",8,"Research & Report","AnythingReality: Robust Online Gaussian Splatting SLAM for Open-Vocabulary VR Scene Exploration","AnythingReality proposes an integrated framework combining robust online 3D Gaussian splatting SLAM, real-time VR exploration, and speech-driven vision-language model interaction. Instead of relying on clean depth or external poses, the system fuses ORB-SLAM3 pose estimation with online Gaussian reconstruction to handle noisy real-world inputs. A VR pipeline supports immersive incremental reconstructions, while a semantic module transcribes voice commands, produces scene descriptions, and logs points of interest. Experiments show improved quality over state-of-the-art online Gaussian splatting methods and an 88% VLM object-recognition rate.","AnythingReality: Robust Online Gaussian Splatting SLAM for Open-Vocabulary VR Scene Exploration  \nTimofei Kozlov 1 * Dmitrii Maliukov 1† Andrey Marchenko2‡ Miguel Altamirano Cabrera 1§  \nDzmitry Tsetserukou 1¶  \narXiv :2607 .09260v1 [ cs .CV] 10 Jul 2026  \n1 Intelligent Space Robotics Laboratory, Skolkovo Institute of Science and Technology, Moscow, Moscow, Russian Federation  \n2 NLP Research Center, Moscow, Russian Federation  \nFigure 1: Renders of a 3D GS reconstruction with 3 state-of-the-art methods and with our integrated system. Flow chart on the right illustrates the semantic interaction pipeline of our project.  \nABSTRACT  \nWe present a novel integrated architecture for robust online 3D Gaussian splatting, real-time VR exploration, and speech-driven Vision-Language-Model interaction. Unlike methods assuming clean depth or external poses, our system combines ORB-SLAM3-based pose estimation with online Gaussian reconstruction for noisy real-world data. A VR pipeline enables immersive exploration of incremental reconstructions; a semantic module transcribes voice commands, generates scene descriptions, and records points of interest. Against state-of-the-art online Gaussian splatting methods, we improve image quality on our dataset (+14.5% PSNR,+8 .6% SSIM, −14 .3% LPIPS) and TUM-RGBD (+11 .7% PSNR,+7.8% SSIM, −21 .6% LPIPS), with comparable or superior frame rates via quality–speed configurations. We achieve 88% VLM object-recognition rate.  \nIndex Terms: 3D Gaussian splatting, online reconstruction, virtual reality, orb-slam  \n1 INTRODUCTION  \nRecent advances in neural scene representations have significantly improved the quality and efficiency of 3D reconstruction. In particular, Gaussian Splatting has emerged as a powerful approach for real-time rendering of complex scenes, making it especially attractive for interactive visualization and immersive applications.  \nHowever, most self-sufficient online reconstruction pipelines require accurate depth measurements for tracking, which can be difficult to obtain in real-world settings. [1] These pipelines often rely  \n* [e-mail: Timofei.Kozlov@skoltech.ru](e-mail: Timofei.Kozlov@skoltech.ru)[ ](e-mail: Timofei.Kozlov@skoltech.ru)†[e-mail: Dmitrii.Maliukov@skoltech.ru](e-mail: Dmitrii.Maliukov@skoltech.ru)[ ](e-mail: Dmitrii.Maliukov@skoltech.ru)‡[e-mail: Timofei Kozlov@skoltech.ru](e-mail: Timofei Kozlov@skoltech.ru)  \n§[e-mail: m.altamirano@skoltech.ru](e-mail: m.altamirano@skoltech.ru)[ ](e-mail: m.altamirano@skoltech.ru)¶e-mail: d.tsetserukou@skoltech.ru  \non ICP-based tracking methods, which can be sensitive to missing depth and noisy measurements of the standard RGB-D cameras. Some aspects of the inaccuracies in measurements can be avoided by using RGB-D cameras that utilize TOF sensors. However, this type of camera is significantly less budget-friendly and has problems operating under direct sunlight and is less suitable for mobile platforms in terms of energy and compute. Other online reconstruction pipelines depend on external pose estimators, such as mobile robotic systems with reliable odometry and SLAM systems. [2] Moreover, most reconstruction pipelines still provide limited support for direct user interaction during the reconstruction process. As a result, users often remain passive observers, with few tools for inspecting, querying, or semantically understanding the evolving scene representation.  \nThese limitations suggest that online reconstruction should not only focus on producing accurate 3D representations, but also on making them accessible, interpretable, and interactive for human users. In this context, Virtual Reality has become an effective interface for exploring reconstructed 3D environments, as it allows users to observe spatial structure, geometry, and visual appearance from an egocentric, human-scale perspective [3, 4, 5] . This immersive form of inspection is particularly useful when reconstructed scenes need to be assessed not only as geometr","cbCaip7ytZ1pR5EL","https://ap.wps.com/l/cbCaip7ytZ1pR5EL","pdf",5189932,3,1,4,"English","en",105,"# Abstract\n# Introduction\n# System Architecture","[{\"question\":\"How does AnythingReality achieve robust online 3D reconstruction without relying on accurate depth measurements?\",\"answer\":\"It uses ORB-SLAM3-based pose estimation to reduce dependence on clean depth-based tracking, then integrates the pose stream with an online Gaussian reconstruction module for noisy data.\"},{\"question\":\"What interactive features are provided in the VR exploration pipeline?\",\"answer\":\"Users can explore scenes during reconstruction, ask natural-language questions about the current Gaussian-rendered view, and create open-vocabulary points of interest via voice commands.\"},{\"question\":\"What improvements does the system report versus state-of-the-art online Gaussian splatting methods?\",\"answer\":\"The approach improves image quality on its dataset (+14.5% PSNR, +8.6% SSIM, −14.3% LPIPS) and on TUM-RGBD (+11.7% PSNR, +7.8% SSIM, −21.6% LPIPS), while maintaining comparable or better frame rates through quality–speed configurations.\"}]",1784179322,10,{"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":85,"head_meta":87,"extra_data":89,"updated_unix":28},"anythingreality-robust-online-gaussian-splatting-slam-for-open-vocabulary-vr-scene-exploration","",{"@graph":36,"@context":84},[37,52,67],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"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":20},"https://docshare.wps.com/document/research-report/",{"item":51,"name":13,"@type":43,"position":22},"https://docshare.wps.com/document/anythingreality-robust-online-gaussian-splatting-slam-for-open-vocabulary-vr-scene-exploration/82273/",{"url":51,"name":13,"@type":53,"author":54,"headline":13,"publisher":56,"fileFormat":59,"inLanguage":24,"description":14,"dateModified":60,"datePublished":61,"encodingFormat":59,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":55},"Person",{"url":41,"name":57,"@type":58},"DocShare","Organization","application/pdf","2026-07-21","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":20},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"How does AnythingReality achieve robust online 3D reconstruction without relying on accurate depth measurements?","Question",{"text":74,"@type":75},"It uses ORB-SLAM3-based pose estimation to reduce dependence on clean depth-based tracking, then integrates the pose stream with an online Gaussian reconstruction module for noisy data.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"What interactive features are provided in the VR exploration pipeline?",{"text":79,"@type":75},"Users can explore scenes during reconstruction, ask natural-language questions about the current Gaussian-rendered view, and create open-vocabulary points of interest via voice commands.",{"name":81,"@type":72,"acceptedAnswer":82},"What improvements does the system report versus state-of-the-art online Gaussian splatting methods?",{"text":83,"@type":75},"The approach improves image quality on its dataset (+14.5% PSNR, +8.6% SSIM, −14.3% LPIPS) and on TUM-RGBD (+11.7% PSNR, +7.8% SSIM, −21.6% LPIPS), while maintaining comparable or better frame rates through quality–speed configurations.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":57,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,119,122,127,130,133],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":22,"doc_module":4,"doc_module_name":46,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":46,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":29,"doc_module":4,"doc_module_name":46,"category_name":131,"show_sort_weight":29,"slug":132},"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":46,"category_name":135,"show_sort_weight":105,"slug":136},19,"General","general"]