[{"data":1,"prerenderedAt":629},["ShallowReactive",2],{"method-huang2024_2dgs":3},{"method":4,"reference":50,"equipment":72,"figures":81,"results":122},{"id":5,"label":6,"shortName":7,"title":8,"year":9,"era":10,"cluster":11,"scope":12,"keyIdeaZh":13,"keyIdeaEn":14,"fulltextStatus":15,"publicationStatus":16,"recommendation":17,"constructionRelevance":18,"validationEnvironment":19,"strengths":21,"limitations":26,"sensors":32,"platform":34,"estimator":35,"association":36,"timeModel":37,"deskew":37,"loopClosure":38,"globalOptimization":38,"mapRepresentation":39,"prior":40,"outputGeometry":41,"compute":42,"codeUrl":43,"codeLicense":44,"relatedVersions":45},"huang2024_2dgs","Huang et al., 2024a","2DGS","2D Gaussian Splatting for Geometrically Accurate Radiance Fields",2024,"recent","C09","map_representation_or_reconstruction","2DGS 將三維體積壓縮為一組有方向的二維平面高斯圓盤，使基元在多視角下具一致的幾何，並加入深度失真與法向一致性正則化。網格以渲染深度圖經 TSDF 融合取得。作者在 DTU 以 Chamfer 距離、在 Tanks and Temples 以 F1 評估幾何，指出幾何與影像品質之間存在取捨。","Replaces 3D Gaussians with planar 2D Gaussian disks plus depth-distortion and normal regularization to obtain view-consistent geometry, meshed by TSDF fusion of rendered depth.","full_text_reviewed","peer_reviewed_published","background","原論文未涉及營建；2DGS 已被用於隧道與建築外殼的營建研究，見 [qian2026_tunnel2dgs]、[chowdhury2026_gema]（依其摘要）。",[20],"public_benchmark",[22,23,24,25],"Lowest mean DTU Chamfer distance among compared methods: 0.80 (30k iterations) vs NeuS 0.84, VolSDF 0.86, SuGaR 1.33 and 3DGS 1.96 (Table 1)","About 100 times faster than SDF baselines on DTU (10.9 min vs more than 12 h) (Table 1)","Tanks and Temples mean F1 0.32 vs 0.09 for 3DGS and 0.19 for SuGaR (Table 2)","Competitive novel-view synthesis on Mip-NeRF 360 (outdoor PSNR 24.34 vs 24.64 for 3DGS) (Table 4)",[27,28,29,30,31],"Assumes fully opaque surfaces; semi-transparent surfaces such as glass are problematic (Limitations, Fig. 12)","Densification favours texture-rich over geometry-rich areas (Limitations)","Regularization trades image quality against geometry and may over-smooth (Limitations)","Tends to create holes in areas with high light intensity (Fig. 12)","Tanks and Temples mean F1 (0.32) stays below implicit SDF methods Neuralangelo (0.50) and NeuS (0.38) (Table 2)",[33],"monocular camera (multi-view images)",[],"not_applicable (per-scene optimization; poses from COLMAP)","direct photometric loss with depth-distortion and normal-consistency regularization","not_applicable","none","2D oriented planar Gaussian disks (surfels) with perspective-correct ray-splat intersection","COLMAP cameras and sparse points","mesh by TSDF fusion (Open3D, voxel size 0.004, truncation 0.02) of rendered median depth maps of the training views; median depth outperforms expected depth and screened Poisson reconstruction in the DTU ablation","Custom CUDA kernels built on the 3DGS framework; all experiments on a single RTX 3090 GPU (written 'GTX RTX3090'); DTU training 5.5 min (15k iterations) or 10.9 min (30k), Tanks and Temples 15.5 min, DTU model 52 MB","https:\u002F\u002Fgithub.com\u002Fhbb1\u002F2d-gaussian-splatting","Gaussian-Splatting License (non-commercial research use per LICENSE.md)",[46],{"relation":47,"title":48,"doi_or_url":49},"preprint","arXiv:2403.17888 (comment: corrected Eq. 7)","https:\u002F\u002Farxiv.org\u002Fabs\u002F2403.17888",{"id":5,"kind":51,"shortName":7,"title":8,"authors":52,"year":9,"venue":58,"venueType":59,"publisher":60,"volumeIssuePages":61,"doi":62,"arxivId":63,"url":64,"firstPublicDate":65,"publicationStatus":16,"metadataStatus":66,"fulltextStatus":15,"era":10,"classicReason":37,"codeUrl":43,"cluster":11,"topics":67,"mdpi":68,"verification":69,"label":6,"fulltextRoute":70,"versionRead":71,"addedByCensus":68},"method",[53,54,55,56,57],"Binbin Huang","Zehao Yu","Anpei Chen","Andreas Geiger","Shenghua Gao","Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers (SIGGRAPH 2024)","conference","ACM","pp. 1-11","10.1145\u002F3641519.3657428","2403.17888","https:\u002F\u002Fapi.crossref.org\u002Fworks\u002F10.1145\u002F3641519.3657428","2024-03-26","metadata_verified",[11],false,"confirmed","arXiv","arXiv v3 HTML (2025-02-22, comment 'Corrected Eq.7'), which carries the SIGGRAPH 2024 Conference Papers metadata; ACM version of record (CC BY 4.0) not opened",[73],{"category":74,"model":75,"canonical":76,"role":77,"dataset":78,"specs":79,"locator":80},"compute","GTX RTX3090 (as written)","GTX RTX3090","compute for runtime",null,"single GPU for all experiments","Sec. 6.1",[82,95,106,114],{"refId":5,"refLabel":6,"fig":83,"whatZh":84,"license":85,"licenseUrl":86,"sourceUrl":87,"src":88,"width":89,"height":90,"thumb":91,"thumbWidth":92,"thumbHeight":93,"modified":94},"Fig. 1","2DGS 以二維定向圓盤重建真實場景，並輸出一致的法向、深度與無雜訊網格","CC BY 4.0 (arXiv v3; ACM version also CC BY 4.0)","http:\u002F\u002Fcreativecommons.org\u002Flicenses\u002Fby\u002F4.0\u002F","https:\u002F\u002Farxiv.org\u002Fhtml\u002F2403.17888v3\u002Fteaser.png","\u002Ffigure-files\u002Fhuang2024_2dgs\u002Ffig-1.webp",1364,549,"\u002Ffigure-files\u002Fhuang2024_2dgs\u002Ffig-1.thumb.webp",480,193,"converted to WebP",{"refId":5,"refLabel":6,"fig":96,"whatZh":97,"license":98,"licenseUrl":86,"sourceUrl":99,"src":100,"width":101,"height":102,"thumb":103,"thumbWidth":92,"thumbHeight":104,"modified":105},"Fig. 2","3DGS 與 2DGS 的比較：3DGS 在不同視角使用不同交會平面，2DGS 具多視角一致性","CC BY 4.0","https:\u002F\u002Farxiv.org\u002Fhtml\u002F2403.17888v3\u002Ffigures\u002Fteaser2dgs.png","\u002Ffigure-files\u002Fhuang2024_2dgs\u002Ffig-2.webp",1400,463,"\u002Ffigure-files\u002Fhuang2024_2dgs\u002Ffig-2.thumb.webp",159,"resized to at most 1400 px wide and converted to WebP",{"refId":5,"refLabel":6,"fig":107,"whatZh":108,"license":98,"licenseUrl":86,"sourceUrl":109,"src":110,"width":101,"height":111,"thumb":112,"thumbWidth":92,"thumbHeight":113,"modified":105},"Fig. 5","DTU 資料集的定性比較，2DGS 表面細緻且雜訊少","https:\u002F\u002Farxiv.org\u002Fhtml\u002F2403.17888v3\u002Ffigures\u002Fdtu.png","\u002Ffigure-files\u002Fhuang2024_2dgs\u002Ffig-5.webp",615,"\u002Ffigure-files\u002Fhuang2024_2dgs\u002Ffig-5.thumb.webp",211,{"refId":5,"refLabel":6,"fig":115,"whatZh":116,"license":98,"licenseUrl":86,"sourceUrl":117,"src":118,"width":101,"height":119,"thumb":120,"thumbWidth":92,"thumbHeight":121,"modified":105},"Fig. 10","Tanks and Temples 資料集的表面重建定性結果","https:\u002F\u002Farxiv.org\u002Fhtml\u002F2403.17888v3\u002Fsupp-tnt.png","\u002Ffigure-files\u002Fhuang2024_2dgs\u002Ffig-10.webp",704,"\u002Ffigure-files\u002Fhuang2024_2dgs\u002Ffig-10.thumb.webp",241,{"totalRows":123,"groupCount":124,"groups":125,"others":628},47,4,[126,425,552,603],{"slug":127,"group":128,"sourceId":5,"sourceLabel":6,"table":129,"selfRows":130,"metrics":131,"seqs":140,"entrants":177,"cells":195,"outcomes":416,"locators":419,"hardware":420,"wordings":422,"notes":423},"huang2024-2dgs-table-1","huang2024_2dgs:Table 1","Table 1",34,[132,135,137],{"label":133,"unit":134,"statistic":134,"alignment":134},"Chamfer distance (CD)","not_reported",{"label":133,"unit":134,"statistic":136,"alignment":134},"mean",{"label":138,"unit":139,"statistic":134,"alignment":38},"Time (training\u002Freconstruction time)","min",[141,145,147,149,151,153,155,157,159,161,163,165,167,169,171,173,175],{"dataset":142,"sequence":143,"environment":144},"DTU","scan 24","DTU: 15 scenes with 49 or 69 images at 1600 x 1200, downsampled to 800 x 600 (scene type not described in the paper)",{"dataset":142,"sequence":146,"environment":144},"scan 37",{"dataset":142,"sequence":148,"environment":144},"scan 40",{"dataset":142,"sequence":150,"environment":144},"scan 55",{"dataset":142,"sequence":152,"environment":144},"scan 63",{"dataset":142,"sequence":154,"environment":144},"scan 65",{"dataset":142,"sequence":156,"environment":144},"scan 69",{"dataset":142,"sequence":158,"environment":144},"scan 83",{"dataset":142,"sequence":160,"environment":144},"scan 97",{"dataset":142,"sequence":162,"environment":144},"scan 105",{"dataset":142,"sequence":164,"environment":144},"scan 106",{"dataset":142,"sequence":166,"environment":144},"scan 110",{"dataset":142,"sequence":168,"environment":144},"scan 114",{"dataset":142,"sequence":170,"environment":144},"scan 118",{"dataset":142,"sequence":172,"environment":144},"scan 122",{"dataset":142,"sequence":174,"environment":144},"Mean (15 scans)",{"dataset":142,"sequence":176,"environment":144},"all 15 scans",[178,182,184,186,189,191,193],{"name":179,"methodId":180,"linkable":181,"proposed":68,"self":68},"NeRF (Mildenhall et al., 2021)","nerf2020",true,{"name":183,"methodId":78,"linkable":68,"proposed":68,"self":68},"VolSDF (Yariv et al., 2021)",{"name":185,"methodId":78,"linkable":68,"proposed":68,"self":68},"NeuS (Wang et al., 2021)",{"name":187,"methodId":188,"linkable":181,"proposed":68,"self":68},"3DGS (Kerbl et al., 2023)","kerbl2023_3dgs",{"name":190,"methodId":78,"linkable":68,"proposed":68,"self":68},"SuGaR (Guédon and Lepetit, 2023)",{"name":192,"methodId":5,"linkable":181,"proposed":181,"self":181},"2DGS-15k (Ours)",{"name":194,"methodId":5,"linkable":181,"proposed":181,"self":181},"2DGS-30k (Ours)",[196,200,203,206,209,211,214,217,220,223,226,229,232,235,238,241,244,246,248,250,252,254,256,258,260,262,264,265,267,269,271,273,275,277,278,279,281,283,285,287,289,291,293,295,297,299,301,303,304,306,308,309,311,313,315,317,319,321,323,325,327,329,331,333,334,336,338,340,342,343,345,347,348,350,352,353,355,357,358,360,362,364,365,366,367,368,370,372,373,375,377,378,379,381,382,384,385,386,387,388,390,391,393,394,396,398,399,401,402,403,404,405,406,407,409,410,411,412,414],[197,197,197,198,199,197,199,199,197],0,1.9,-1,[197,197,201,202,199,197,199,199,197],1,1.6,[197,197,204,205,199,197,199,199,197],2,1.85,[197,197,207,208,199,197,199,199,197],3,0.58,[197,197,124,210,199,197,199,199,197],2.28,[197,197,212,213,199,197,199,199,197],5,1.27,[197,197,215,216,199,197,199,199,197],6,1.47,[197,197,218,219,199,197,199,199,197],7,1.67,[197,197,221,222,199,197,199,199,197],8,2.05,[197,197,224,225,199,197,199,199,197],9,1.07,[197,197,227,228,199,197,199,199,197],10,0.88,[197,197,230,231,199,197,199,199,197],11,2.53,[197,197,233,234,199,197,199,199,197],12,1.06,[197,197,236,237,199,197,199,199,197],13,1.15,[197,197,239,240,199,197,199,199,197],14,0.96,[197,201,242,243,199,197,199,199,197],15,1.49,[197,204,245,78,197,197,197,199,197],16,[201,197,197,247,199,197,199,199,197],1.14,[201,197,201,249,199,197,199,199,197],1.26,[201,197,204,251,199,197,199,199,197],0.81,[201,197,207,253,199,197,199,199,197],0.49,[201,197,124,255,199,197,199,199,197],1.25,[201,197,212,257,199,197,199,199,197],0.7,[201,197,215,259,199,197,199,199,197],0.72,[201,197,218,261,199,197,199,199,197],1.29,[201,197,221,263,199,197,199,199,197],1.18,[201,197,224,257,199,197,199,199,197],[201,197,227,266,199,197,199,199,197],0.66,[201,197,230,268,199,197,199,199,197],1.08,[201,197,233,270,199,197,199,199,197],0.42,[201,197,236,272,199,197,199,199,197],0.61,[201,197,239,274,199,197,199,199,197],0.55,[201,201,242,276,199,197,199,199,197],0.86,[201,204,245,78,197,197,197,199,197],[204,197,197,201,199,197,199,199,197],[204,197,201,280,199,197,199,199,197],1.37,[204,197,204,282,199,197,199,199,197],0.93,[204,197,207,284,199,197,199,199,197],0.43,[204,197,124,286,199,197,199,199,197],1.1,[204,197,212,288,199,197,199,199,197],0.65,[204,197,215,290,199,197,199,199,197],0.57,[204,197,218,292,199,197,199,199,197],1.48,[204,197,221,294,199,197,199,199,197],1.09,[204,197,224,296,199,197,199,199,197],0.83,[204,197,227,298,199,197,199,199,197],0.52,[204,197,230,300,199,197,199,199,197],1.2,[204,197,233,302,199,197,199,199,197],0.35,[204,197,236,253,199,197,199,199,197],[204,197,239,305,199,197,199,199,197],0.54,[204,201,242,307,199,197,199,199,197],0.84,[204,204,245,78,197,197,197,199,197],[207,197,197,310,199,197,199,199,197],2.14,[207,197,201,312,199,197,199,199,197],1.53,[207,197,204,314,199,197,199,199,197],2.08,[207,197,207,316,199,197,199,199,197],1.68,[207,197,124,318,199,197,199,199,197],3.49,[207,197,212,320,199,197,199,199,197],2.21,[207,197,215,322,199,197,199,199,197],1.43,[207,197,218,324,199,197,199,199,197],2.07,[207,197,221,326,199,197,199,199,197],2.22,[207,197,224,328,199,197,199,199,197],1.75,[207,197,227,330,199,197,199,199,197],1.79,[207,197,230,332,199,197,199,199,197],2.55,[207,197,233,312,199,197,199,199,197],[207,197,236,335,199,197,199,199,197],1.52,[207,197,239,337,199,197,199,199,197],1.5,[207,201,242,339,199,197,199,199,197],1.96,[207,204,245,341,199,197,197,199,197],11.2,[124,197,197,216,199,197,199,199,197],[124,197,201,344,199,197,199,199,197],1.33,[124,197,204,346,199,197,199,199,197],1.13,[124,197,207,272,199,197,199,199,197],[124,197,124,349,199,197,199,199,197],2.25,[124,197,212,351,199,197,199,199,197],1.71,[124,197,215,237,199,197,199,199,197],[124,197,218,354,199,197,199,199,197],1.63,[124,197,221,356,199,197,199,199,197],1.62,[124,197,224,225,199,197,199,199,197],[124,197,227,359,199,197,199,199,197],0.79,[124,197,230,361,199,197,199,199,197],2.45,[124,197,233,363,199,197,199,199,197],0.98,[124,197,236,228,199,197,199,199,197],[124,197,239,359,199,197,199,199,197],[124,201,242,344,199,197,199,199,197],[124,204,245,78,201,197,197,199,197],[212,197,197,369,199,197,199,199,197],0.48,[212,197,201,371,199,197,199,199,197],0.92,[212,197,204,270,199,197,199,199,197],[212,197,207,374,199,197,199,199,197],0.4,[212,197,124,376,199,197,199,199,197],1.04,[212,197,212,296,199,197,199,199,197],[212,197,215,296,199,197,199,199,197],[212,197,218,380,199,197,199,199,197],1.36,[212,197,221,213,199,197,199,199,197],[212,197,224,383,199,197,199,199,197],0.76,[212,197,227,259,199,197,199,199,197],[212,197,230,354,199,197,199,199,197],[212,197,233,374,199,197,199,199,197],[212,197,236,383,199,197,199,199,197],[212,197,239,389,199,197,199,199,197],0.6,[212,201,242,296,199,197,199,199,197],[212,204,245,392,199,197,197,199,197],5.5,[215,197,197,369,199,197,199,199,197],[215,197,201,395,199,197,199,199,197],0.91,[215,197,204,397,199,197,199,199,197],0.39,[215,197,207,397,199,197,199,199,197],[215,197,124,400,199,197,199,199,197],1.01,[215,197,212,296,199,197,199,199,197],[215,197,215,251,199,197,199,199,197],[215,197,218,380,199,197,199,199,197],[215,197,221,213,199,197,199,199,197],[215,197,224,383,199,197,199,199,197],[215,197,227,257,199,197,199,199,197],[215,197,230,408,199,197,199,199,197],1.4,[215,197,233,374,199,197,199,199,197],[215,197,236,383,199,197,199,199,197],[215,197,239,298,199,197,199,199,197],[215,201,242,413,199,197,199,199,197],0.8,[215,204,245,415,199,197,197,199,197],10.9,[417,418],"reported as '>12h' (not a point value)","reported as '~1h' (not a point value)",[129],[421],"2DGS experiments on a single RTX 3090 (written 'GTX RTX3090'); hardware for baselines not stated",[],[424],"Chamfer distance per DTU scan (15 scans) and mean; meshes of 3DGS and 2DGS by TSDF fusion of rendered depth; unit not stated in the paper; images downsampled to 800 x 600; COLMAP sparse points for initialization",{"slug":426,"group":427,"sourceId":5,"sourceLabel":6,"table":428,"selfRows":221,"metrics":429,"seqs":435,"entrants":454,"cells":467,"outcomes":544,"locators":547,"hardware":548,"wordings":549,"notes":550},"huang2024-2dgs-table-2","huang2024_2dgs:Table 2","Table 2",[430,432,433],{"label":431,"unit":38,"statistic":134,"alignment":134},"F1 score",{"label":431,"unit":38,"statistic":136,"alignment":134},{"label":434,"unit":139,"statistic":134,"alignment":38},"Time (training time)",[436,440,442,444,446,448,450,452],{"dataset":437,"sequence":438,"environment":439},"Tanks and Temples","Barn","Tanks and Temples scenes Barn, Caterpillar, Courthouse, Ignatius, Meetingroom and Truck (environment type not described in the paper)",{"dataset":437,"sequence":441,"environment":439},"Caterpillar",{"dataset":437,"sequence":443,"environment":439},"Courthouse",{"dataset":437,"sequence":445,"environment":439},"Ignatius",{"dataset":437,"sequence":447,"environment":439},"Meetingroom",{"dataset":437,"sequence":449,"environment":439},"Truck",{"dataset":437,"sequence":451,"environment":439},"Mean (6 scenes)",{"dataset":437,"sequence":453,"environment":439},"all 6 scenes",[455,457,459,461,463,465],{"name":456,"methodId":78,"linkable":68,"proposed":68,"self":68},"NeuS",{"name":458,"methodId":78,"linkable":68,"proposed":68,"self":68},"Geo-Neus",{"name":460,"methodId":78,"linkable":68,"proposed":68,"self":68},"Neurlangelo (as written; Neuralangelo)",{"name":462,"methodId":78,"linkable":68,"proposed":68,"self":68},"SuGaR",{"name":464,"methodId":188,"linkable":181,"proposed":68,"self":68},"3DGS",{"name":466,"methodId":5,"linkable":181,"proposed":181,"self":181},"Ours",[468,470,471,473,474,476,478,480,481,483,485,487,488,490,491,492,493,494,496,498,500,502,503,505,506,508,510,512,513,515,516,518,519,521,522,524,526,528,529,530,532,534,536,537,539,540,541,542],[197,197,197,469,199,197,199,199,197],0.29,[197,197,201,469,199,197,199,199,197],[197,197,204,472,199,197,199,199,197],0.17,[197,197,207,296,199,197,199,199,197],[197,197,124,475,199,197,199,199,197],0.24,[197,197,212,477,199,197,199,199,197],0.45,[197,201,215,479,199,197,199,199,197],0.38,[197,204,218,78,197,197,197,199,197],[201,197,197,482,199,197,199,199,197],0.33,[201,197,201,484,199,197,199,199,197],0.26,[201,197,204,486,199,197,199,199,197],0.12,[201,197,207,259,199,197,199,199,197],[201,197,124,489,199,197,199,199,197],0.2,[201,197,212,477,199,197,199,199,197],[201,201,215,302,199,197,199,199,197],[201,204,218,78,197,197,197,199,197],[204,197,197,257,199,197,199,199,197],[204,197,201,495,199,197,199,199,197],0.36,[204,197,204,497,199,197,199,199,197],0.28,[204,197,207,499,199,197,199,199,197],0.89,[204,197,124,501,199,197,199,199,197],0.32,[204,197,212,369,199,197,199,199,197],[204,201,215,504,199,197,199,199,197],0.5,[204,204,218,78,197,197,197,199,197],[207,197,197,507,199,197,199,199,197],0.14,[207,197,201,509,199,197,199,199,197],0.16,[207,197,204,511,199,197,199,199,197],0.08,[207,197,207,482,199,197,199,199,197],[207,197,124,514,199,197,199,199,197],0.15,[207,197,212,484,199,197,199,199,197],[207,201,215,517,199,197,199,199,197],0.19,[207,204,218,78,201,197,197,199,197],[124,197,197,520,199,197,199,199,197],0.13,[124,197,201,511,199,197,199,199,197],[124,197,204,523,199,197,199,199,197],0.09,[124,197,207,525,199,197,199,199,197],0.04,[124,197,124,527,199,197,199,199,197],0.01,[124,197,212,517,199,197,199,199,197],[124,201,215,523,199,197,199,199,197],[124,204,218,531,199,197,197,199,197],14.3,[212,197,197,533,199,197,199,199,197],0.41,[212,197,201,535,199,197,199,199,197],0.23,[212,197,204,509,199,197,199,199,197],[212,197,207,538,199,197,199,199,197],0.51,[212,197,124,472,199,197,199,199,197],[212,197,212,477,199,197,199,199,197],[212,201,215,501,199,197,199,199,197],[212,204,218,543,199,197,197,199,197],15.5,[545,546],"reported as '>24h' (not a point value)","reported as '>1h' (not a point value)",[428],[421],[],[551],"F1 score per Tanks and Temples scene and mean; distance threshold not stated in the paper (TnT benchmark protocol)",{"slug":553,"group":554,"sourceId":555,"sourceLabel":556,"table":557,"selfRows":124,"metrics":558,"seqs":567,"entrants":572,"cells":579,"outcomes":596,"locators":597,"hardware":598,"wordings":600,"notes":601},"qian2026-tunnel2dgs-table-3","qian2026_tunnel2dgs:Table 3","qian2026_tunnel2dgs","Qian et al., 2026","Table 3",[559,561,563,565],{"label":560,"unit":139,"statistic":134,"alignment":38},"Modeling time, Foundation: SfM (converted from h min s)",{"label":562,"unit":139,"statistic":134,"alignment":38},"Modeling time, Total (converted from h min s)",{"label":564,"unit":139,"statistic":134,"alignment":38},"Modeling time, Step 1: 2DGS (no mask) (converted from h min s)",{"label":566,"unit":139,"statistic":134,"alignment":38},"Modeling time, Step 2: TSDF (converted from h min s)",[568],{"dataset":569,"sequence":570,"environment":571},"Zhejiang shield tunnel UAV video (authors)","80 m section, 210 images","under-construction shield tunnel",[573,575,577],{"name":574,"methodId":78,"linkable":68,"proposed":68,"self":68},"MVS + surface reconstruction (No. 1)",{"name":576,"methodId":5,"linkable":181,"proposed":68,"self":181},"2DGS (no mask) + TSDF (No. 2)",{"name":578,"methodId":555,"linkable":68,"proposed":181,"self":68},"2DGS (mask) + TSDF (No. 3)",[580,582,584,585,587,589,591,592,594],[197,197,197,581,199,197,197,199,197],12.12,[197,201,197,583,199,197,197,199,197],104.15,[201,197,197,581,199,197,197,199,197],[201,204,197,586,199,197,197,199,197],51,[201,207,197,588,199,197,197,199,197],4.47,[201,201,197,590,199,197,197,199,197],68.58,[204,197,197,581,199,197,197,199,197],[204,207,197,593,199,197,197,199,197],3.43,[204,201,197,595,199,197,197,199,197],65.55,[],[557],[599],"Intel Core i9-14900KF, NVIDIA GeForce RTX 4080, 64 GB RAM, Windows 11",[],[602],"Modeling time per pipeline stage for the three mesh models on the same workstation; MVS dense stage run with GPU acceleration; face counts of Nos. 2 and 3 matched to No. 1 (about 10 million) via TSDF voxel size",{"slug":604,"group":605,"sourceId":555,"sourceLabel":556,"table":606,"selfRows":201,"metrics":607,"seqs":611,"entrants":614,"cells":617,"outcomes":621,"locators":622,"hardware":624,"wordings":625,"notes":626},"qian2026-tunnel2dgs-text-sec-geometric-accuracy-evaluation","qian2026_tunnel2dgs:Text Sec. Geometric Accuracy Evaluation","Text Sec. Geometric Accuracy Evaluation",[608],{"label":609,"unit":610,"statistic":136,"alignment":134},"Average point distance to model No. 1","u (arbitrary, no absolute scale)",[612],{"dataset":569,"sequence":613,"environment":571},"80 m section",[615,616],{"name":576,"methodId":5,"linkable":181,"proposed":68,"self":181},{"name":578,"methodId":555,"linkable":68,"proposed":181,"self":68},[618,619],[197,197,197,523,199,197,199,199,197],[201,197,197,620,199,197,199,199,197],0.083,[],[623],"Experimental Results, Geometric Accuracy Evaluation; Fig. 14 text",[],[],[627],"Mean point-to-point distance from uniformly sampled mesh points of each 2DGS+TSDF model to the MVS+surface-reconstruction mesh (No. 1, designated reference); models have no absolute scale (unit u); registration method not stated",[],1790510664561]