[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-86102-en":3,"doc-seo-86102-105":30,"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":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},86102,1374391974468,"Eden","https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0",8,"Research & Report","AsySplat Efficient Asymmetric 3D Gaussian Splatting for Long Sequence Scene Modeling","Recent generalizable 3D Gaussian Splatting advances long-sequence novel view synthesis (NVS) but incurs heavy redundant computation. The work argues redundancy can be reduced because high-precision geometry is not strictly required for rendering quality, while appearance learning is generally easier than geometry recovery. It introduces an asymmetric architecture that decouples geometry and appearance: a coarse-grained geometry branch and a fine-grained appearance branch, linked via bilateral connections. Experiments on 32-view 960P inputs match optimization-based methods with nearly 800× speedup and fewer parameters overall, improving training and inference efficiency.","arXiv :2607 . 10995v1 [ cs .CV] 13 Jul 2026  \nAsySplat: Efficient Asymmetric 3D Gaussian Splatting for Long-Sequence Scene Modeling  \nYingji Zhong 1 Dave Zhenyu Chen2 Fuzhao Ou3 Youyu Chen2 Zhihao Li2 Lanqing Hong2 Dan Xu 1  \n1HKUST 2Huawei Noah’s Ark Lab 3 CityU  \nAbstract  \nRecent generalizable 3D Gaussian Splatting models have advanced long-sequence novel view synthesis (NVS), but at the cost of substantial redundant computation.  \nWe identify that the redundancy can be mitigated based on two observations:(i) high-precision geometry is not strictly required for high-quality NVS; (ii) appearance learning is generally easier than geometry recovery. Motivated by these insights, we propose an asymmetric architecture that decouples geometry and appearance modeling. The geometry branch processes coarse-grained tokens with most of the parameters for multi-view reconstruction, while the appearance branch operates on fine-grained tokens to capture details using significantly fewer parameters. The two branches interact through bilateral connections, enabling mutual guidance for their respective tasks. This task-aware asymmetry reduces the computational redundancy and allocates the computation more judiciously, thereby increasing parameter efficiency and enabling smaller models to achieve strong performance. On 32-view 960P inputs, our model matches optimizationbased methods while delivering nearly 800× speedup, and surpasses the zeroshot performance of state-of-the-art generalizable models with markedly fewer parameters and reduced training/inference overhead, achieving an overall efficiency improvement. The project page is at [https://zhongyingji.github.io/asysplat/](https://zhongyingji.github.io/asysplat/) .  \n1 Introduction  \n3D Gaussian Splatting (3DGS) [30, 59, 37] has emerged as an effective representation for novel view synthesis (NVS) . Recent research has shifted toward generalizable 3DGS models [5, 14, 8, 43, 54, 63, 52], also referred to as Gaussian Reconstruction Models, which predict 3D Gaussians from networks, bypassing per-scene optimization through data-driven priors. While most methods focus on sparse inputs, real-world captures often consist of long sequences, making such a setting more practical. With the rise of high-resolution devices and the growing expectations of visual detail, it becomes increasingly important to extend generalizable 3DGS methods to handle high-resolution, long-sequence inputs [71] .  \nMost current methods like LongLRM [71] stack numerous layers [46, 20] to predict depth and Gaussian appearance attributes directly from patchified tokens. This design concentrates all computation on inter-frame token interactions. This costly process stems from the need for small patch sizes to achieve high-quality rendering. This raises a fundamental question: Does such simplistic stacking introduce computational redundancy that drives up parameter demand? In this work, we explore whether a more deliberate allocation of computation can improve parameter efficiency in NVS, thereby empowering compact models to achieve competitive performance.  \nIn this work, we highlight two observations that are crucial for computation allocation but overlooked in prior generalizable 3DGS methods: (i) High-precision geometry is not a must for high-quality rendering. While geometry matters, the high-fidelity 3DGS rendering largely arises from the alpha compositing of Gaussians during splatting, rather than from highly precise geometry. This is consistent with reports of high rendering quality despite modest performance in surface reconstruc  \nPreprint.  \nFigure 1: Our AsySplat reduces computational redundancy in the long-sequence Gaussian reconstruction model, by decoupling it into asymmetric geometry and appearance branches. The asymmetry spans token granularity, computation allocation, and parameter distribution, which is motivated by two observations: (i) High-precision geometry is not a must for high-quality rendering. (ii) ","cbCaimmmeCQsaj2J","https://ap.wps.com/l/cbCaimmmeCQsaj2J","pdf",5816101,4,1,17,"English","en",105,"# Introduction\n## Background on generalizable 3D Gaussian Splatting\n## Motivation: computational redundancy in long sequences\n## Core observations and proposed asymmetric design","[{\"question\":\"What problem does AsySplat address in long-sequence novel view synthesis?\",\"answer\":\"It targets the substantial redundant computation introduced by existing generalizable 3D Gaussian Splatting designs when processing long sequences.\"},{\"question\":\"Why does AsySplat decouple geometry and appearance modeling?\",\"answer\":\"It leverages two observations: high-precision geometry is not strictly necessary for high-quality rendering, and appearance attributes are generally easier to learn than geometry recovery.\"},{\"question\":\"How does AsySplat allocate computation between the two branches?\",\"answer\":\"The geometry branch uses coarse-grained tokens with larger patch sizes and receives most parameters for reconstruction, while the appearance branch uses fine-grained tokens with smaller patch sizes and fewer parameters to capture details.\"}]",1784208528,43,{"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":86,"head_meta":88,"extra_data":90,"updated_unix":28},"asysplat-efficient-asymmetric-3d-gaussian-splatting-for-long-sequence-scene-modeling","",{"@graph":36,"@context":85},[37,53,68],{"@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":20},"https://docshare.wps.com/document/asysplat-efficient-asymmetric-3d-gaussian-splatting-for-long-sequence-scene-modeling/86102/",{"url":52,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-26","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 problem does AsySplat address in long-sequence novel view synthesis?","Question",{"text":75,"@type":76},"It targets the substantial redundant computation introduced by existing generalizable 3D Gaussian Splatting designs when processing long sequences.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Why does AsySplat decouple geometry and appearance modeling?",{"text":80,"@type":76},"It leverages two observations: high-precision geometry is not strictly necessary for high-quality rendering, and appearance attributes are generally easier to learn than geometry recovery.",{"name":82,"@type":73,"acceptedAnswer":83},"How does AsySplat allocate computation between the two branches?",{"text":84,"@type":76},"The geometry branch uses coarse-grained tokens with larger patch sizes and receives most parameters for reconstruction, while the appearance branch uses fine-grained tokens with smaller patch sizes and fewer parameters to capture details.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"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":106,"slug":138},19,"General","general"]