[{"data":1,"prerenderedAt":761},["ShallowReactive",2],{"method-qin2023geotransformer":3},{"method":4,"reference":58,"equipment":83,"figures":91,"results":92},{"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":22,"limitations":27,"sensors":32,"platform":34,"estimator":36,"association":37,"timeModel":38,"deskew":38,"loopClosure":39,"globalOptimization":40,"mapRepresentation":41,"prior":42,"outputGeometry":43,"compute":44,"codeUrl":45,"codeLicense":46,"relatedVersions":47},"qin2023geotransformer","Qin et al., 2023","GeoTransformer","GeoTransformer: Fast and Robust Point Cloud Registration With Geometric Transformer",2023,"recent","C02","registration_component","GeoTransformer 屬學習式、免關鍵點的配準：先在降採樣的超點（superpoint）間比對，再傳播到稠密點。其幾何 Transformer 編碼點對距離與三點角度，使特徵對剛體變換不變，並在低重疊情形下保持穩健；摘要指出匹配精度高到不需 RANSAC 即可估計轉換。","Keypoint-free learned registration matching superpoints with a transformation-invariant geometric transformer, removing the need for RANSAC.","full_text_reviewed","peer_reviewed_published","supplementary","not_reported",[20,21],"public_benchmark","simulation",[23,24,25,26],"improves inlier ratio by 18 to 31 points and registration recall by over 7 points on 3DLoMatch (abstract, Sec. 4.2)","RANSAC-free LGR: registration recall 91.8% (3DMatch) and 74.5% (3DLoMatch) with 0.088 s average total time vs 1.633 s with RANSAC-50k (Table 2)","KITTI: RTE 6.8 cm, RRE 0.24 deg, RR 99.8% with LGR (Table 3)","Augmented ICL-NUIM with the 3DMatch-trained model: mean ATE 14.12 cm, best on three of four trajectories (Table 5)",[28,29,30,31],"learned priors may not transfer across sensors or scenes (inference; cross-dataset degradation reported for a different learned method in lim2025kissmatcher Fig. 5)","uniformly downsampled superpoints can cause a large memory footprint and computational cost for inputs covering a large area; an extra downsampling stage was needed on KITTI and Augmented ICL-NUIM (Sec. 5)","uniform superpoint sampling is inflexible and can split a single object into several patches (Sec. 5)","the backbone is not invariant to rotation, so performance drops under large rotations (Sec. 4.4)",[33],"none of its own; public benchmarks: 3DMatch and 3DLoMatch indoor RGB-D scene fragments, KITTI odometry LiDAR scans (pairs at least 10 m apart, ground truth refined by ICP), ModelNet40 synthetic CAD points, Augmented ICL-NUIM synthetic RGB-D with a noise model, 4DMatch and 4DLoMatch non-rigid animations",[35],"not_applicable (offline benchmark evaluation)","local-to-global registration (LGR): weighted SVD on the point correspondences of each superpoint match gives candidate transforms, the candidate with most inliers within an acceptance radius is kept and re-estimated on inliers for Nr = 5 iterations; RANSAC-50k and plain weighted SVD are also evaluated","KPConv-FPN backbone; geometric self-attention (pair-wise distance and triplet-wise angle embeddings) interleaved three times with feature-based cross-attention; Gaussian correlation with dual normalisation selects the top Nc superpoint matches (256 at test time); an optimal-transport layer (Sinkhorn) with mutual top-k extracts dense point correspondences inside matched patches","not_applicable","none","none for pairwise registration; in the Augmented ICL-NUIM multiway test, pairwise GeoTransformer results are followed by global pose-graph optimisation","superpoints and dense points","trained model per benchmark (40 to 200 epochs); the Augmented ICL-NUIM test reuses the 3DMatch model without fine-tuning","rigid transformation","PyTorch on an RTX 3090 GPU; 3DMatch\u002F3DLoMatch average total time 0.088 s per pair with LGR vs 1.633 s with RANSAC-50k (pose time 0.013 s vs 1.558 s); the lite model runs at 0.073 s, about 13 fps","https:\u002F\u002Fgithub.com\u002Fqinzheng93\u002FGeoTransformer","MIT (LICENSE file checked)",[48,52,56],{"relation":49,"title":50,"doi_or_url":51},"conference_version","Geometric Transformer for Fast and Robust Point Cloud Registration (CVPR 2022, pp. 11133-11142)","10.1109\u002FCVPR52688.2022.01086",{"relation":53,"title":54,"doi_or_url":55},"preprint","arXiv:2308.03768 (TPAMI version, 2023-07-25); CVPR version arXiv:2202.06688 per arXiv comment","https:\u002F\u002Farxiv.org\u002Fabs\u002F2308.03768",{"relation":57,"title":7,"doi_or_url":45},"code_release",{"id":5,"kind":59,"shortName":7,"title":8,"authors":60,"year":9,"venue":69,"venueType":70,"publisher":71,"volumeIssuePages":72,"doi":73,"arxivId":74,"url":75,"firstPublicDate":76,"publicationStatus":16,"metadataStatus":77,"fulltextStatus":15,"era":10,"classicReason":38,"codeUrl":45,"cluster":11,"topics":78,"mdpi":79,"verification":80,"label":6,"fulltextRoute":81,"versionRead":82,"addedByCensus":79},"method",[61,62,63,64,65,66,67,68],"Zheng Qin","Hao Yu","Changjian Wang","Yulan Guo","Yuxing Peng","Slobodan Ilic","Dewen Hu","Kai Xu","IEEE Transactions on Pattern Analysis and Machine Intelligence","journal","IEEE","45(8):9806-9821","10.1109\u002Ftpami.2023.3259038","2308.03768","https:\u002F\u002Fapi.crossref.org\u002Fworks\u002F10.1109\u002Ftpami.2023.3259038","2023-03-20","metadata_verified",[11],false,"confirmed","arXiv","arXiv 2308.03768v1 (2023-07-25), the TPAMI-accepted extended version per the arXiv comment; IEEE version of record not compared",[84],{"category":85,"model":86,"canonical":86,"role":87,"dataset":88,"specs":89,"locator":90},"compute","RTX 3090 GPU","compute for runtime",null,"used for training and testing in PyTorch","Sec. 4.1",[],{"totalRows":93,"groupCount":94,"groups":95,"others":750},70,6,[96,339,570,646],{"slug":97,"group":98,"sourceId":5,"sourceLabel":6,"table":99,"selfRows":100,"metrics":101,"seqs":113,"entrants":124,"cells":161,"outcomes":332,"locators":333,"hardware":334,"wordings":336,"notes":337},"qin2023geotransformer-table-2","qin2023geotransformer:Table 2","Table 2",30,[102,105,109,111],{"label":103,"unit":104,"statistic":18,"alignment":38},"Registration Recall RR (%)","%",{"label":106,"unit":107,"statistic":108,"alignment":38},"Model time (feature extraction)","s","mean",{"label":110,"unit":107,"statistic":108,"alignment":38},"Pose time (transformation estimation)",{"label":112,"unit":107,"statistic":108,"alignment":38},"Total time",[114,118,121],{"dataset":115,"sequence":116,"environment":117},"3DMatch","test split","indoor RGB-D scene fragments",{"dataset":119,"sequence":116,"environment":120},"3DLoMatch","indoor RGB-D scene fragments (low overlap)",{"dataset":122,"sequence":123,"environment":117},"3DMatch and 3DLoMatch","averaged over all point cloud pairs",[125,129,131,133,135,137,139,141,143,145,147,149,151,153,155,157,159],{"name":126,"methodId":127,"linkable":128,"proposed":79,"self":79},"FCGF + RANSAC-50k (5000 samples)","choy2019fcgf",true,{"name":130,"methodId":88,"linkable":79,"proposed":79,"self":79},"D3Feat + RANSAC-50k (5000 samples)",{"name":132,"methodId":88,"linkable":79,"proposed":79,"self":79},"SpinNet + RANSAC-50k (5000 samples)",{"name":134,"methodId":88,"linkable":79,"proposed":79,"self":79},"Predator + RANSAC-50k (5000 samples)",{"name":136,"methodId":88,"linkable":79,"proposed":79,"self":79},"CoFiNet + RANSAC-50k (5000 samples)",{"name":138,"methodId":5,"linkable":128,"proposed":128,"self":128},"GeoTransformer (ours) + RANSAC-50k (5000 samples)",{"name":140,"methodId":5,"linkable":128,"proposed":128,"self":128},"GeoTransformer lite (ours, shared geometric self-attention) + RANSAC-50k (5000 samples)",{"name":142,"methodId":127,"linkable":128,"proposed":79,"self":79},"FCGF + weighted SVD (250 samples)",{"name":144,"methodId":88,"linkable":79,"proposed":79,"self":79},"D3Feat + weighted SVD (250 samples)",{"name":146,"methodId":88,"linkable":79,"proposed":79,"self":79},"SpinNet + weighted SVD (250 samples)",{"name":148,"methodId":88,"linkable":79,"proposed":79,"self":79},"Predator + weighted SVD (250 samples)",{"name":150,"methodId":88,"linkable":79,"proposed":79,"self":79},"CoFiNet + weighted SVD (250 samples)",{"name":152,"methodId":5,"linkable":128,"proposed":128,"self":128},"GeoTransformer (ours) + weighted SVD (250 samples)",{"name":154,"methodId":5,"linkable":128,"proposed":128,"self":128},"GeoTransformer lite (ours, shared geometric self-attention) + weighted SVD (250 samples)",{"name":156,"methodId":88,"linkable":79,"proposed":79,"self":79},"CoFiNet + LGR (all samples)",{"name":158,"methodId":5,"linkable":128,"proposed":128,"self":128},"GeoTransformer (ours) + LGR (all samples)",{"name":160,"methodId":5,"linkable":128,"proposed":128,"self":128},"GeoTransformer lite (ours, shared geometric self-attention) + LGR (all samples)",[162,166,169,172,174,177,179,181,183,185,187,189,191,193,195,197,199,200,202,204,206,209,211,213,215,217,220,222,224,226,228,230,232,234,236,238,241,243,244,246,248,251,253,254,255,256,259,261,262,264,266,269,271,272,274,276,279,281,282,284,286,289,291,292,293,295,298,300,301,302,304,307,309,310,312,314,317,319,320,322,324,326,328,329,330],[163,163,163,164,165,163,165,165,163],0,85.1,-1,[163,163,167,168,165,163,165,165,163],1,40.1,[163,167,170,171,165,163,163,165,163],2,0.052,[163,170,170,173,165,163,163,165,163],3.326,[163,175,170,176,165,163,163,165,163],3,3.378,[167,163,163,178,165,163,165,165,163],81.6,[167,163,167,180,165,163,165,165,163],37.2,[167,167,170,182,165,163,163,165,163],0.024,[167,170,170,184,165,163,163,165,163],3.088,[167,175,170,186,165,163,163,165,163],3.112,[170,163,163,188,165,163,165,165,163],88.6,[170,163,167,190,165,163,165,165,163],59.8,[170,167,170,192,165,163,163,165,163],60.248,[170,170,170,194,165,163,163,165,163],0.388,[170,175,170,196,165,163,163,165,163],60.636,[175,163,163,198,165,163,165,165,163],89,[175,163,167,190,165,163,165,165,163],[175,167,170,201,165,163,163,165,163],0.032,[175,170,170,203,165,163,163,165,163],5.12,[175,175,170,205,165,163,163,165,163],5.152,[207,163,163,208,165,163,165,165,163],4,89.3,[207,163,167,210,165,163,165,165,163],67.5,[207,167,170,212,165,163,163,165,163],0.115,[207,170,170,214,165,163,163,165,163],1.807,[207,175,170,216,165,163,163,165,163],1.922,[218,163,163,219,165,163,165,165,163],5,92.3,[218,163,167,221,165,163,165,165,163],75.4,[218,167,170,223,165,163,163,165,163],0.075,[218,170,170,225,165,163,163,165,163],1.558,[218,175,170,227,165,163,163,165,163],1.633,[94,163,163,229,165,163,165,165,163],92.2,[94,163,167,231,165,163,165,165,163],74.9,[94,167,170,233,165,163,163,165,163],0.06,[94,170,170,235,165,163,163,165,163],1.546,[94,175,170,237,165,163,163,165,163],1.606,[239,163,163,240,165,163,165,165,163],7,42.1,[239,163,167,242,165,163,165,165,163],3.9,[239,167,170,171,165,163,163,165,163],[239,170,170,245,165,163,163,165,163],0.008,[239,175,170,247,165,163,163,165,163],0.056,[249,163,163,250,165,163,165,165,163],8,37.4,[249,163,167,252,165,163,165,165,163],2.8,[249,167,170,182,165,163,163,165,163],[249,170,170,245,165,163,163,165,163],[249,175,170,201,165,163,163,165,163],[257,163,163,258,165,163,165,165,163],9,34,[257,163,167,260,165,163,165,165,163],2.5,[257,167,170,192,165,163,163,165,163],[257,170,170,263,165,163,163,165,163],0.006,[257,175,170,265,165,163,163,165,163],60.254,[267,163,163,268,165,163,165,165,163],10,50,[267,163,167,270,165,163,165,165,163],6.4,[267,167,170,201,165,163,163,165,163],[267,170,170,273,165,163,163,165,163],0.009,[267,175,170,275,165,163,163,165,163],0.041,[277,163,163,278,165,163,165,165,163],11,64.6,[277,163,167,280,165,163,165,165,163],21.6,[277,167,170,212,165,163,163,165,163],[277,170,170,283,165,163,163,165,163],0.003,[277,175,170,285,165,163,163,165,163],0.118,[287,163,163,288,165,163,165,165,163],12,86.7,[287,163,167,290,165,163,165,165,163],60.5,[287,167,170,223,165,163,163,165,163],[287,170,170,283,165,163,163,165,163],[287,175,170,294,165,163,163,165,163],0.078,[296,163,163,297,165,163,165,165,163],13,87.5,[296,163,167,299,165,163,165,165,163],61.4,[296,167,170,233,165,163,163,165,163],[296,170,170,283,165,163,163,165,163],[296,175,170,303,165,163,163,165,163],0.063,[305,163,163,306,165,163,165,165,163],14,87.6,[305,163,167,308,165,163,165,165,163],64.8,[305,167,170,212,165,163,163,165,163],[305,170,170,311,165,163,163,165,163],0.028,[305,175,170,313,165,163,163,165,163],0.143,[315,163,163,316,165,163,165,165,163],15,91.8,[315,163,167,318,165,163,165,165,163],74.5,[315,167,170,223,165,163,163,165,163],[315,170,170,321,165,163,163,165,163],0.013,[315,175,170,323,165,163,163,165,163],0.088,[325,163,163,316,165,163,165,165,163],16,[325,163,167,327,165,163,165,165,163],74.2,[325,167,170,233,165,163,163,165,163],[325,170,170,321,165,163,163,165,163],[325,175,170,331,165,163,163,165,163],0.073,[],[99],[335],"RTX 3090 GPU (Sec. 4.1); CPU not stated",[],[338],"3DMatch (overlap above 30%) and 3DLoMatch (10% to 30%) test pairs; registration recall = share of pairs with transformation RMSE below 0.2 m; model time = feature extraction, pose time = transformation estimation, averaged over all pairs",{"slug":340,"group":341,"sourceId":342,"sourceLabel":343,"table":344,"selfRows":345,"metrics":346,"seqs":353,"entrants":374,"cells":385,"outcomes":564,"locators":565,"hardware":566,"wordings":567,"notes":568},"sun2025nss-table-5","sun2025nss:Table 5","sun2025nss","Sun et al., 2025","Table 5",18,[347,349],{"label":348,"unit":104,"statistic":18,"alignment":39},"registration recall (RRE \u003C 10 deg and RTE \u003C 0.2 m)",{"label":350,"unit":351,"statistic":352,"alignment":39},"RMSE [m] (column in Table 5; definition not given in the evaluation-metrics text)","m","RMSE",[354,358,360,362,364,366,368,370,372],{"dataset":355,"sequence":356,"environment":357},"Nothing Stands Still (NSS)","Cross-Area split, all spatiotemporal pairs","indoor building areas under construction or renovation (NSS areas A-F)",{"dataset":355,"sequence":359,"environment":357},"Cross-Stage split, all spatiotemporal pairs",{"dataset":355,"sequence":361,"environment":357},"Original split, all spatiotemporal pairs",{"dataset":355,"sequence":363,"environment":357},"Cross-Area split, only same-stage pairs",{"dataset":355,"sequence":365,"environment":357},"Cross-Stage split, only same-stage pairs",{"dataset":355,"sequence":367,"environment":357},"Original split, only same-stage pairs",{"dataset":355,"sequence":369,"environment":357},"Cross-Area split, only different-stage pairs",{"dataset":355,"sequence":371,"environment":357},"Cross-Stage split, only different-stage pairs",{"dataset":355,"sequence":373,"environment":357},"Original split, only different-stage pairs",[375,378,380,382,384],{"name":376,"methodId":377,"linkable":128,"proposed":79,"self":79},"FPFH","rusu2009fpfh",{"name":379,"methodId":127,"linkable":128,"proposed":79,"self":79},"FCGF",{"name":381,"methodId":88,"linkable":79,"proposed":79,"self":79},"D3Feat",{"name":383,"methodId":88,"linkable":79,"proposed":79,"self":79},"Predator",{"name":7,"methodId":5,"linkable":128,"proposed":79,"self":128},[386,388,390,392,394,396,398,400,402,404,406,408,410,412,414,416,418,420,422,424,426,428,430,432,434,436,438,440,442,444,446,448,450,452,454,456,458,460,462,464,466,468,469,471,473,475,477,479,481,483,485,487,489,491,493,495,497,499,501,503,505,507,509,511,513,515,517,519,521,523,525,527,529,531,533,535,537,539,541,543,545,547,549,551,553,555,556,558,560,562],[163,163,163,387,165,163,165,165,163],22.83,[163,167,163,389,165,163,165,165,163],3.3,[163,163,167,391,165,163,165,165,163],18.73,[163,167,167,393,165,163,165,165,163],2.53,[163,163,170,395,165,163,165,165,163],11.7,[163,167,170,397,165,163,165,165,163],2.52,[167,163,163,399,165,163,165,165,163],28.22,[167,167,163,401,165,163,165,165,163],2.07,[167,163,167,403,165,163,165,165,163],37.7,[167,167,167,405,165,163,165,165,163],1.81,[167,163,170,407,165,163,165,165,163],24.43,[167,167,170,409,165,163,165,165,163],2.24,[170,163,163,411,165,163,165,165,163],31.77,[170,167,163,413,165,163,165,165,163],1.98,[170,163,167,415,165,163,165,165,163],51.37,[170,167,167,417,165,163,165,165,163],1.62,[170,163,170,419,165,163,165,165,163],22.73,[170,167,170,421,165,163,165,165,163],2.37,[175,163,163,423,165,163,165,165,163],55.53,[175,167,163,425,165,163,165,165,163],1.09,[175,163,167,427,165,163,165,165,163],76.73,[175,167,167,429,165,163,165,165,163],0.77,[175,163,170,431,165,163,165,165,163],64.97,[175,167,170,433,165,163,165,165,163],0.71,[207,163,163,435,165,163,165,165,163],38.13,[207,167,163,437,165,163,165,165,163],1.24,[207,163,167,439,165,163,165,165,163],47.78,[207,167,167,441,165,163,165,165,163],0.98,[207,163,170,443,165,163,165,165,163],39.07,[207,167,170,445,165,163,165,165,163],0.96,[163,163,175,447,165,163,165,165,163],32.86,[163,167,175,449,165,163,165,165,163],2.46,[163,163,207,451,165,163,165,165,163],46.4,[163,167,207,453,165,163,165,165,163],1.94,[163,163,218,455,165,163,165,165,163],30.82,[163,167,218,457,165,163,165,165,163],2.58,[167,163,175,459,165,163,165,165,163],39.32,[167,167,175,461,165,163,165,165,163],1.88,[167,163,207,463,165,163,165,165,163],44.65,[167,167,207,465,165,163,165,165,163],1.77,[167,163,218,467,165,163,165,165,163],42.86,[167,167,218,409,165,163,165,165,163],[170,163,175,470,165,163,165,165,163],43.62,[170,167,175,472,165,163,165,165,163],1.91,[170,163,207,474,165,163,165,165,163],58.47,[170,167,207,476,165,163,165,165,163],1.48,[170,163,218,478,165,163,165,165,163],36.51,[170,167,218,480,165,163,165,165,163],2.09,[175,163,175,482,165,163,165,165,163],76.8,[175,167,175,484,165,163,165,165,163],0.81,[175,163,207,486,165,163,165,165,163],87.49,[175,167,207,488,165,163,165,165,163],0.44,[175,163,218,490,165,163,165,165,163],92.99,[175,167,218,492,165,163,165,165,163],0.27,[207,163,175,494,165,163,165,165,163],50.88,[207,167,175,496,165,163,165,165,163],1.07,[207,163,207,498,165,163,165,165,163],54.07,[207,167,207,500,165,163,165,165,163],0.79,[207,163,218,502,165,163,165,165,163],55.59,[207,167,218,504,165,163,165,165,163],0.69,[163,163,94,506,165,163,165,165,163],1.06,[163,167,94,508,165,163,165,165,163],4.88,[163,163,239,510,165,163,165,165,163],0.82,[163,167,239,512,165,163,165,165,163],4.23,[163,163,249,514,165,163,165,165,163],0.42,[163,167,249,516,165,163,165,165,163],4.21,[167,163,94,518,165,163,165,165,163],5.21,[167,167,94,520,165,163,165,165,163],3.22,[167,163,239,522,165,163,165,165,163],14.06,[167,167,239,524,165,163,165,165,163],4.15,[167,163,249,526,165,163,165,165,163],10.52,[167,167,249,528,165,163,165,165,163],3.28,[170,163,94,530,165,163,165,165,163],6.12,[170,167,94,532,165,163,165,165,163],2.01,[170,163,239,534,165,163,165,165,163],12.85,[170,167,239,536,165,163,165,165,163],2.4,[170,163,249,538,165,163,165,165,163],4.76,[170,167,249,540,165,163,165,165,163],2.75,[175,163,94,542,165,163,165,165,163],9.49,[175,167,94,544,165,163,165,165,163],1.71,[175,163,239,546,165,163,165,165,163],18.42,[175,167,239,548,165,163,165,165,163],2.03,[175,163,249,550,165,163,165,165,163],28.42,[175,167,249,552,165,163,165,165,163],1.28,[207,163,94,554,165,163,165,165,163],10.55,[207,167,94,417,165,163,165,165,163],[207,163,239,557,165,163,165,165,163],13.39,[207,167,239,559,165,163,165,165,163],2.25,[207,163,249,561,165,163,165,165,163],17.51,[207,167,249,563,165,163,165,165,163],1.31,[],[344],[],[],[569],"Pairwise spatiotemporal registration on NSS; success = RRE \u003C 10 deg and RTE \u003C 0.2 m; methods retrained per split following original protocols. TE and RE columns (successful pairs \u002F all pairs) not transcribed.",{"slug":571,"group":572,"sourceId":342,"sourceLabel":343,"table":573,"selfRows":249,"metrics":574,"seqs":577,"entrants":592,"cells":596,"outcomes":640,"locators":641,"hardware":642,"wordings":643,"notes":644},"sun2025nss-table-6","sun2025nss:Table 6","Table 6",[575],{"label":576,"unit":104,"statistic":18,"alignment":39},"registration recall",[578,581,583,585,587,589,590,591],{"dataset":115,"sequence":579,"environment":580},"Standard overlap (30%+)","indoor RGB-D reconstructions (3DMatch benchmark)",{"dataset":582,"sequence":579,"environment":357},"NSS (all), original split",{"dataset":584,"sequence":579,"environment":357},"NSS (same-stage only), original split",{"dataset":586,"sequence":579,"environment":357},"NSS (different-stage only), original split",{"dataset":119,"sequence":588,"environment":580},"Low overlap (10-30%)",{"dataset":582,"sequence":588,"environment":357},{"dataset":584,"sequence":588,"environment":357},{"dataset":586,"sequence":588,"environment":357},[593,594,595],{"name":381,"methodId":88,"linkable":79,"proposed":79,"self":79},{"name":383,"methodId":88,"linkable":79,"proposed":79,"self":79},{"name":7,"methodId":5,"linkable":128,"proposed":79,"self":128},[597,599,600,602,604,606,608,610,612,614,616,618,620,621,622,624,626,628,630,631,633,635,637,639],[163,163,163,598,165,163,165,165,163],82.2,[167,163,163,198,165,163,165,165,163],[170,163,163,601,165,163,165,165,163],92,[163,163,167,603,165,163,165,165,163],34.1,[167,163,167,605,165,163,165,165,163],58.8,[170,163,167,607,165,163,165,165,163],40.3,[163,163,170,609,165,163,165,165,163],47.9,[167,163,170,611,165,163,165,165,163],83.1,[170,163,170,613,165,163,165,165,163],54.3,[163,163,175,615,165,163,165,165,163],6.6,[167,163,175,617,165,163,165,165,163],10.6,[170,163,175,619,165,163,165,165,163],12.5,[163,163,207,180,165,163,165,165,163],[167,163,207,190,165,163,165,165,163],[170,163,207,623,165,163,165,165,163],75,[163,163,218,625,165,163,165,165,163],25.2,[167,163,218,627,165,163,165,165,163],47.6,[170,163,218,629,165,163,165,165,163],32.6,[163,163,94,629,165,163,165,165,163],[167,163,94,632,165,163,165,165,163],62.4,[170,163,94,634,165,163,165,165,163],42.2,[163,163,239,636,165,163,165,165,163],4.4,[167,163,239,638,165,163,165,165,163],5.9,[170,163,239,638,165,163,165,165,163],[],[573],[],[],[645],"Registration recall of the three best methods on NSS (original split) versus 3DMatch and 3DLoMatch; benchmark values for 3DMatch and 3DLoMatch as listed by the authors.",{"slug":647,"group":648,"sourceId":5,"sourceLabel":6,"table":649,"selfRows":94,"metrics":650,"seqs":658,"entrants":663,"cells":683,"outcomes":741,"locators":745,"hardware":746,"wordings":747,"notes":748},"qin2023geotransformer-table-3","qin2023geotransformer:Table 3","Table 3",[651,654,657],{"label":652,"unit":653,"statistic":18,"alignment":38},"Relative Translation Error RTE (cm)","cm",{"label":655,"unit":656,"statistic":18,"alignment":38},"Relative Rotation Error RRE (deg)","deg",{"label":103,"unit":104,"statistic":18,"alignment":38},[659],{"dataset":660,"sequence":661,"environment":662},"KITTI odometry","sequences 8-10","outdoor driving LiDAR scans",[664,666,667,668,670,671,673,675,677,679,681],{"name":665,"methodId":88,"linkable":79,"proposed":79,"self":79},"3DFeat-Net",{"name":379,"methodId":127,"linkable":128,"proposed":79,"self":79},{"name":381,"methodId":88,"linkable":79,"proposed":79,"self":79},{"name":669,"methodId":88,"linkable":79,"proposed":79,"self":79},"SpinNet",{"name":383,"methodId":88,"linkable":79,"proposed":79,"self":79},{"name":672,"methodId":88,"linkable":79,"proposed":79,"self":79},"CoFiNet",{"name":674,"methodId":5,"linkable":128,"proposed":128,"self":128},"GeoTransformer (ours, RANSAC-50k)",{"name":676,"methodId":88,"linkable":79,"proposed":79,"self":79},"FMR",{"name":678,"methodId":88,"linkable":79,"proposed":79,"self":79},"DGR",{"name":680,"methodId":88,"linkable":79,"proposed":79,"self":79},"HRegNet",{"name":682,"methodId":5,"linkable":128,"proposed":128,"self":128},"GeoTransformer (ours, LGR)",[684,686,688,690,692,694,696,698,699,701,703,705,707,709,710,711,713,715,716,718,719,720,722,724,726,728,730,732,733,735,737,738,740],[163,163,163,685,165,163,165,165,163],25.9,[163,167,163,687,165,163,165,165,163],0.25,[163,170,163,689,165,163,165,165,163],96,[167,163,163,691,165,163,165,165,163],9.5,[167,167,163,693,165,163,165,165,163],0.3,[167,170,163,695,165,163,165,165,163],96.6,[170,163,163,697,165,163,165,165,163],7.2,[170,167,163,693,165,163,165,165,163],[170,170,163,700,165,163,165,165,163],99.8,[175,163,163,702,165,163,165,165,163],9.9,[175,167,163,704,165,163,165,165,163],0.47,[175,170,163,706,165,163,165,165,163],99.1,[207,163,163,708,165,163,165,165,163],6.8,[207,167,163,492,165,163,165,165,163],[207,170,163,700,165,163,165,165,163],[218,163,163,712,165,163,165,165,163],8.2,[218,167,163,714,165,163,165,165,163],0.41,[218,170,163,700,165,163,165,165,163],[94,163,163,717,165,163,165,165,163],7.4,[94,167,163,492,165,163,165,165,163],[94,170,163,700,165,163,165,165,163],[239,163,163,721,163,163,165,165,163],66,[239,167,163,723,165,163,165,165,163],1.49,[239,170,163,725,165,163,165,165,163],90.6,[249,163,163,727,167,163,165,165,163],32,[249,167,163,729,165,163,165,165,163],0.37,[249,170,163,731,165,163,165,165,163],98.7,[257,163,163,287,170,163,165,165,163],[257,167,163,734,165,163,165,165,163],0.29,[257,170,163,736,165,163,165,165,163],99.7,[267,163,163,708,165,163,165,165,163],[267,167,163,739,165,163,165,165,163],0.24,[267,170,163,700,165,163,165,165,163],[742,743,744],"approximate ('~66' in table)","approximate ('~32' in table)","approximate ('~12' in table)",[649],[],[],[749],"KITTI odometry sequences 8 to 10 for testing, pairs at least 10 m apart, ground truth refined with ICP; RR = share of pairs with RRE below 5 deg and RTE below 2 m; top block RANSAC-based, bottom block RANSAC-free",[751,756],{"group":752,"slug":753,"sourceLabel":6,"table":344,"selfRows":218,"datasets":754},"qin2023geotransformer:Table 5","qin2023geotransformer-table-5",[755],"Augmented ICL-NUIM",{"group":757,"slug":758,"sourceLabel":343,"table":759,"selfRows":175,"datasets":760},"sun2025nss:Table 9","sun2025nss-table-9","Table 9",[355],1790510664522]