[{"data":1,"prerenderedAt":799},["ShallowReactive",2],{"method-choy2019fcgf":3},{"method":4,"reference":50,"equipment":70,"figures":88,"results":89},{"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":33,"platform":36,"estimator":37,"association":38,"timeModel":39,"deskew":39,"loopClosure":40,"globalOptimization":40,"mapRepresentation":41,"prior":42,"outputGeometry":43,"compute":44,"codeUrl":45,"codeLicense":46,"relatedVersions":47},"choy2019fcgf","Choy et al., 2019","FCGF","Fully Convolutional Geometric Features",2019,"recent","C02","registration_component","FCGF 以 Minkowski Engine 稀疏卷積構成的 ResUNet，一次計算整片點雲每個體素的 32 維幾何特徵，輸入只用座標與常數特徵，不需法向量或局部區塊（patch）前處理。作者提出最難負樣本對比損失與最難三元組損失，並以雜湊方式濾除錨點附近的假負樣本。3DMatch 上特徵匹配召回率 0.952、配準召回率平均 0.82；KITTI 上以 RANSAC 配準的成功率為 97.83% 至 98.92%。屬學習式先驗，跨感測器與場域的泛化需另行驗證。","Learned dense 32-D geometric features from a 3D fully-convolutional network with metric-learning losses, used for correspondence-based registration.","full_text_reviewed","peer_reviewed_published","supplementary","not_reported",[20],"public_benchmark",[22,23,24,25],"feature-match recall 0.952 (0.953 with rotation augmentation) vs 0.947 for PerfectMatch (Table 1)","average registration recall 0.82 vs 0.71 for PPFNet and 0.40 for FPFH on 3DMatch (Table 5)","KITTI with 20 cm voxels: RTE 4.881 cm, RRE 0.170 deg, success 97.83% vs 3DFeat-Net 25.9 cm, 0.57 deg, 95.97% (Table 6)","rotation invariance learned by augmentation, translation invariance inherent to sparse convolution (Sec. 6.3)",[27,28,29,30,31,32],"learned registration pipelines can degrade on unseen sensor patterns (lim2025kissmatcher Fig. 5, shown for Predator; not FCGF itself)","colour input caused overfitting because the dataset was not large or diverse enough; normals gave no meaningful gain (Sec. 6.1)","hardest-triplet loss is prone to collapse and needs random triplets mixed in (Sec. 4.2, Table 3)","64-D features ran out of memory at 2.5 cm voxels (Table 2)","KITTI evaluation uses pairs at least 10 m apart and indirect RTE and RRE after RANSAC; translation error grows with voxel size (Sec. 6.2, 6.6)","end-to-end registration left to future work (Sec. 7)",[34,35],"indoor 3D scan fragments of the 3DMatch benchmark (sensor not named in the paper)","KITTI odometry LiDAR scans (model not named in the paper)",[],"descriptor only; transformations estimated with RANSAC on FCGF correspondences: RANSAC with early termination for the 3DMatch registration recall (Sec. 6.5) and RANSAC for the KITTI RTE and RRE (Sec. 6.6; early termination is not stated for KITTI)","dense 32-D features from a ResUNet of generalized sparse convolutions (Minkowski Engine) on voxel-downsampled points with 1-vectors as input features; trained with hardest-contrastive or hardest-triplet losses using hash-based filtering of false negatives near anchors; matches found by feature similarity","not_applicable","none","sparse tensor (sparse voxel) representation of point clouds (Sec. 3)","trained model: 3DMatch official split, or KITTI sequences 0 to 5 for training with ICP-refined GPS poses as ground truth; random scaling [0.8, 1.2] and random 3D rotation augmentation","32-dimensional per-point features (abstract)","0.019 ms per feature including preprocessing; about 0.36 s per 3DMatch fragment at 2.5 cm voxel and 0.17 s at 5 cm; Intel i7-6950 10-core 3.0 GHz CPU with Nvidia Titan-X Pascal GPU; about 290x faster than PerfectMatch, 169x than 3DMatch and 42x than PPF-FoldNet","https:\u002F\u002Fgithub.com\u002Fchrischoy\u002FFCGF","MIT (LICENSE file checked)",[48],{"relation":49,"title":7,"doi_or_url":45},"code_release",{"id":5,"kind":51,"shortName":7,"title":8,"authors":52,"year":9,"venue":56,"venueType":57,"publisher":58,"volumeIssuePages":59,"doi":60,"arxivId":61,"url":62,"firstPublicDate":63,"publicationStatus":16,"metadataStatus":64,"fulltextStatus":15,"era":10,"classicReason":39,"codeUrl":45,"cluster":11,"topics":65,"mdpi":66,"verification":67,"label":6,"fulltextRoute":68,"versionRead":69,"addedByCensus":66},"component",[53,54,55],"Christopher Choy","Jaesik Park","Vladlen Koltun","2019 IEEE\u002FCVF International Conference on Computer Vision (ICCV)","conference","IEEE","pp. 8957-8965","10.1109\u002Ficcv.2019.00905",null,"https:\u002F\u002Fapi.crossref.org\u002Fworks\u002F10.1109\u002Ficcv.2019.00905","2019-10","metadata_verified",[11],false,"corrected","other","CVF Open Access version of the ICCV 2019 paper (9 pages, read in full); IEEE Xplore version of record (pp. 8957-8965) opened in Chrome under NTU access to confirm the key values of Tables 1, 5, 6 and the Sec. 6.7 hardware",[71,77,79,82],{"category":72,"model":73,"canonical":73,"role":74,"dataset":61,"specs":75,"locator":76},"compute","Intel i7-6950 (10-core, 3.0 GHz)","compute for runtime","10-core 3.0 GHz","Sec. 6.7",{"category":72,"model":78,"canonical":78,"role":74,"dataset":61,"specs":18,"locator":76},"Nvidia Titan-X Pascal",{"category":72,"model":80,"canonical":80,"role":74,"dataset":61,"specs":81,"locator":76},"Intel i7 8-core 3.2 GHz CPU with Nvidia Titan-X Pascal (used by baselines [7], [6], [36])","8-core 3.2 GHz",{"category":83,"model":84,"canonical":84,"role":85,"dataset":86,"specs":18,"locator":87},"lidar","KITTI LIDAR (model not named in the paper)","dataset sensor","KITTI odometry","Sec. 6.1, Fig. 6",[],{"totalRows":90,"groupCount":91,"groups":92,"others":775},63,8,[93,336,410,638],{"slug":94,"group":95,"sourceId":96,"sourceLabel":97,"table":98,"selfRows":99,"metrics":100,"seqs":108,"entrants":129,"cells":142,"outcomes":330,"locators":331,"hardware":332,"wordings":333,"notes":334},"sun2025nss-table-5","sun2025nss:Table 5","sun2025nss","Sun et al., 2025","Table 5",18,[101,104],{"label":102,"unit":103,"statistic":18,"alignment":40},"registration recall (RRE \u003C 10 deg and RTE \u003C 0.2 m)","%",{"label":105,"unit":106,"statistic":107,"alignment":40},"RMSE [m] (column in Table 5; definition not given in the evaluation-metrics text)","m","RMSE",[109,113,115,117,119,121,123,125,127],{"dataset":110,"sequence":111,"environment":112},"Nothing Stands Still (NSS)","Cross-Area split, all spatiotemporal pairs","indoor building areas under construction or renovation (NSS areas A-F)",{"dataset":110,"sequence":114,"environment":112},"Cross-Stage split, all spatiotemporal pairs",{"dataset":110,"sequence":116,"environment":112},"Original split, all spatiotemporal pairs",{"dataset":110,"sequence":118,"environment":112},"Cross-Area split, only same-stage pairs",{"dataset":110,"sequence":120,"environment":112},"Cross-Stage split, only same-stage pairs",{"dataset":110,"sequence":122,"environment":112},"Original split, only same-stage pairs",{"dataset":110,"sequence":124,"environment":112},"Cross-Area split, only different-stage pairs",{"dataset":110,"sequence":126,"environment":112},"Cross-Stage split, only different-stage pairs",{"dataset":110,"sequence":128,"environment":112},"Original split, only different-stage pairs",[130,134,135,137,139],{"name":131,"methodId":132,"linkable":133,"proposed":66,"self":66},"FPFH","rusu2009fpfh",true,{"name":7,"methodId":5,"linkable":133,"proposed":66,"self":133},{"name":136,"methodId":61,"linkable":66,"proposed":66,"self":66},"D3Feat",{"name":138,"methodId":61,"linkable":66,"proposed":66,"self":66},"Predator",{"name":140,"methodId":141,"linkable":133,"proposed":66,"self":66},"GeoTransformer","qin2023geotransformer",[143,147,150,152,154,157,159,161,163,165,167,169,171,173,175,177,179,181,183,186,188,190,192,194,196,199,201,203,205,207,209,211,213,215,217,220,222,224,226,228,230,232,233,235,237,239,241,243,245,247,249,251,253,255,257,259,261,263,265,267,269,272,274,277,279,281,283,285,287,289,291,293,295,297,299,301,303,305,307,309,311,313,315,317,319,321,322,324,326,328],[144,144,144,145,146,144,146,146,144],0,22.83,-1,[144,148,144,149,146,144,146,146,144],1,3.3,[144,144,148,151,146,144,146,146,144],18.73,[144,148,148,153,146,144,146,146,144],2.53,[144,144,155,156,146,144,146,146,144],2,11.7,[144,148,155,158,146,144,146,146,144],2.52,[148,144,144,160,146,144,146,146,144],28.22,[148,148,144,162,146,144,146,146,144],2.07,[148,144,148,164,146,144,146,146,144],37.7,[148,148,148,166,146,144,146,146,144],1.81,[148,144,155,168,146,144,146,146,144],24.43,[148,148,155,170,146,144,146,146,144],2.24,[155,144,144,172,146,144,146,146,144],31.77,[155,148,144,174,146,144,146,146,144],1.98,[155,144,148,176,146,144,146,146,144],51.37,[155,148,148,178,146,144,146,146,144],1.62,[155,144,155,180,146,144,146,146,144],22.73,[155,148,155,182,146,144,146,146,144],2.37,[184,144,144,185,146,144,146,146,144],3,55.53,[184,148,144,187,146,144,146,146,144],1.09,[184,144,148,189,146,144,146,146,144],76.73,[184,148,148,191,146,144,146,146,144],0.77,[184,144,155,193,146,144,146,146,144],64.97,[184,148,155,195,146,144,146,146,144],0.71,[197,144,144,198,146,144,146,146,144],4,38.13,[197,148,144,200,146,144,146,146,144],1.24,[197,144,148,202,146,144,146,146,144],47.78,[197,148,148,204,146,144,146,146,144],0.98,[197,144,155,206,146,144,146,146,144],39.07,[197,148,155,208,146,144,146,146,144],0.96,[144,144,184,210,146,144,146,146,144],32.86,[144,148,184,212,146,144,146,146,144],2.46,[144,144,197,214,146,144,146,146,144],46.4,[144,148,197,216,146,144,146,146,144],1.94,[144,144,218,219,146,144,146,146,144],5,30.82,[144,148,218,221,146,144,146,146,144],2.58,[148,144,184,223,146,144,146,146,144],39.32,[148,148,184,225,146,144,146,146,144],1.88,[148,144,197,227,146,144,146,146,144],44.65,[148,148,197,229,146,144,146,146,144],1.77,[148,144,218,231,146,144,146,146,144],42.86,[148,148,218,170,146,144,146,146,144],[155,144,184,234,146,144,146,146,144],43.62,[155,148,184,236,146,144,146,146,144],1.91,[155,144,197,238,146,144,146,146,144],58.47,[155,148,197,240,146,144,146,146,144],1.48,[155,144,218,242,146,144,146,146,144],36.51,[155,148,218,244,146,144,146,146,144],2.09,[184,144,184,246,146,144,146,146,144],76.8,[184,148,184,248,146,144,146,146,144],0.81,[184,144,197,250,146,144,146,146,144],87.49,[184,148,197,252,146,144,146,146,144],0.44,[184,144,218,254,146,144,146,146,144],92.99,[184,148,218,256,146,144,146,146,144],0.27,[197,144,184,258,146,144,146,146,144],50.88,[197,148,184,260,146,144,146,146,144],1.07,[197,144,197,262,146,144,146,146,144],54.07,[197,148,197,264,146,144,146,146,144],0.79,[197,144,218,266,146,144,146,146,144],55.59,[197,148,218,268,146,144,146,146,144],0.69,[144,144,270,271,146,144,146,146,144],6,1.06,[144,148,270,273,146,144,146,146,144],4.88,[144,144,275,276,146,144,146,146,144],7,0.82,[144,148,275,278,146,144,146,146,144],4.23,[144,144,91,280,146,144,146,146,144],0.42,[144,148,91,282,146,144,146,146,144],4.21,[148,144,270,284,146,144,146,146,144],5.21,[148,148,270,286,146,144,146,146,144],3.22,[148,144,275,288,146,144,146,146,144],14.06,[148,148,275,290,146,144,146,146,144],4.15,[148,144,91,292,146,144,146,146,144],10.52,[148,148,91,294,146,144,146,146,144],3.28,[155,144,270,296,146,144,146,146,144],6.12,[155,148,270,298,146,144,146,146,144],2.01,[155,144,275,300,146,144,146,146,144],12.85,[155,148,275,302,146,144,146,146,144],2.4,[155,144,91,304,146,144,146,146,144],4.76,[155,148,91,306,146,144,146,146,144],2.75,[184,144,270,308,146,144,146,146,144],9.49,[184,148,270,310,146,144,146,146,144],1.71,[184,144,275,312,146,144,146,146,144],18.42,[184,148,275,314,146,144,146,146,144],2.03,[184,144,91,316,146,144,146,146,144],28.42,[184,148,91,318,146,144,146,146,144],1.28,[197,144,270,320,146,144,146,146,144],10.55,[197,148,270,178,146,144,146,146,144],[197,144,275,323,146,144,146,146,144],13.39,[197,148,275,325,146,144,146,146,144],2.25,[197,144,91,327,146,144,146,146,144],17.51,[197,148,91,329,146,144,146,146,144],1.31,[],[98],[],[],[335],"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":337,"group":338,"sourceId":5,"sourceLabel":6,"table":339,"selfRows":340,"metrics":341,"seqs":351,"entrants":356,"cells":369,"outcomes":404,"locators":405,"hardware":406,"wordings":407,"notes":408},"choy2019fcgf-table-6","choy2019fcgf:Table 6","Table 6",15,[342,346,349],{"label":343,"unit":344,"statistic":345,"alignment":40},"Relative Translation Error (RTE)","cm","mean",{"label":347,"unit":348,"statistic":345,"alignment":40},"Relative Rotation Error (RRE)","deg",{"label":350,"unit":103,"statistic":18,"alignment":40},"Succ. rate",[352],{"dataset":353,"sequence":354,"environment":355},"KITTI odometry (registration pairs)","test split","outdoor driving, LiDAR",[357,359,361,363,365,367],{"name":358,"methodId":61,"linkable":66,"proposed":66,"self":66},"3DFeat [34]",{"name":360,"methodId":5,"linkable":133,"proposed":133,"self":133},"FCGF 20cm",{"name":362,"methodId":5,"linkable":133,"proposed":133,"self":133},"FCGF 25cm",{"name":364,"methodId":5,"linkable":133,"proposed":133,"self":133},"FCGF 30cm",{"name":366,"methodId":5,"linkable":133,"proposed":133,"self":133},"FCGF 35cm",{"name":368,"methodId":5,"linkable":133,"proposed":133,"self":133},"FCGF 40cm",[370,372,374,376,378,380,382,384,386,388,390,392,394,396,398,399,401,403],[144,144,144,371,146,144,146,146,144],25.9,[144,148,144,373,146,144,146,146,144],0.57,[144,155,144,375,146,144,146,146,144],95.97,[148,144,144,377,146,144,146,146,144],4.881,[148,148,144,379,146,144,146,146,144],0.17,[148,155,144,381,146,144,146,146,144],97.83,[155,144,144,383,146,144,146,146,144],6.066,[155,148,144,385,146,144,146,146,144],0.213,[155,155,144,387,146,144,146,146,144],98.56,[184,144,144,389,146,144,146,146,144],6.466,[184,148,144,391,146,144,146,146,144],0.228,[184,155,144,393,146,144,146,146,144],98.92,[197,144,144,395,146,144,146,146,144],6.978,[197,148,144,397,146,144,146,146,144],0.254,[197,155,144,393,146,144,146,146,144],[218,144,144,400,146,144,146,146,144],8.025,[218,148,144,402,146,144,146,146,144],0.273,[218,155,144,393,146,144,146,146,144],[],[339],[],[],[409],"KITTI test pairs (scans at least 10 m apart, ICP-refined GPS ground truth, 555 test pairs); RANSAC on features; success if RTE \u003C 2 m and RRE \u003C 5 deg; FCGF rows differ by downsampling voxel size",{"slug":411,"group":412,"sourceId":141,"sourceLabel":413,"table":414,"selfRows":415,"metrics":416,"seqs":426,"entrants":436,"cells":471,"outcomes":631,"locators":632,"hardware":633,"wordings":635,"notes":636},"qin2023geotransformer-table-2","qin2023geotransformer:Table 2","Qin et al., 2023","Table 2",10,[417,419,422,424],{"label":418,"unit":103,"statistic":18,"alignment":39},"Registration Recall RR (%)",{"label":420,"unit":421,"statistic":345,"alignment":39},"Model time (feature extraction)","s",{"label":423,"unit":421,"statistic":345,"alignment":39},"Pose time (transformation estimation)",{"label":425,"unit":421,"statistic":345,"alignment":39},"Total time",[427,430,433],{"dataset":428,"sequence":354,"environment":429},"3DMatch","indoor RGB-D scene fragments",{"dataset":431,"sequence":354,"environment":432},"3DLoMatch","indoor RGB-D scene fragments (low overlap)",{"dataset":434,"sequence":435,"environment":429},"3DMatch and 3DLoMatch","averaged over all point cloud pairs",[437,439,441,443,445,447,449,451,453,455,457,459,461,463,465,467,469],{"name":438,"methodId":5,"linkable":133,"proposed":66,"self":133},"FCGF + RANSAC-50k (5000 samples)",{"name":440,"methodId":61,"linkable":66,"proposed":66,"self":66},"D3Feat + RANSAC-50k (5000 samples)",{"name":442,"methodId":61,"linkable":66,"proposed":66,"self":66},"SpinNet + RANSAC-50k (5000 samples)",{"name":444,"methodId":61,"linkable":66,"proposed":66,"self":66},"Predator + RANSAC-50k (5000 samples)",{"name":446,"methodId":61,"linkable":66,"proposed":66,"self":66},"CoFiNet + RANSAC-50k (5000 samples)",{"name":448,"methodId":141,"linkable":133,"proposed":133,"self":66},"GeoTransformer (ours) + RANSAC-50k (5000 samples)",{"name":450,"methodId":141,"linkable":133,"proposed":133,"self":66},"GeoTransformer lite (ours, shared geometric self-attention) + RANSAC-50k (5000 samples)",{"name":452,"methodId":5,"linkable":133,"proposed":66,"self":133},"FCGF + weighted SVD (250 samples)",{"name":454,"methodId":61,"linkable":66,"proposed":66,"self":66},"D3Feat + weighted SVD (250 samples)",{"name":456,"methodId":61,"linkable":66,"proposed":66,"self":66},"SpinNet + weighted SVD (250 samples)",{"name":458,"methodId":61,"linkable":66,"proposed":66,"self":66},"Predator + weighted SVD (250 samples)",{"name":460,"methodId":61,"linkable":66,"proposed":66,"self":66},"CoFiNet + weighted SVD (250 samples)",{"name":462,"methodId":141,"linkable":133,"proposed":133,"self":66},"GeoTransformer (ours) + weighted SVD (250 samples)",{"name":464,"methodId":141,"linkable":133,"proposed":133,"self":66},"GeoTransformer lite (ours, shared geometric self-attention) + weighted SVD (250 samples)",{"name":466,"methodId":61,"linkable":66,"proposed":66,"self":66},"CoFiNet + LGR (all samples)",{"name":468,"methodId":141,"linkable":133,"proposed":133,"self":66},"GeoTransformer (ours) + LGR (all samples)",{"name":470,"methodId":141,"linkable":133,"proposed":133,"self":66},"GeoTransformer lite (ours, shared geometric self-attention) + LGR (all samples)",[472,474,476,478,480,482,484,486,488,490,492,494,496,498,500,502,504,505,507,509,511,513,515,517,519,521,523,525,527,529,531,533,535,537,539,541,543,545,546,548,550,552,554,555,556,557,560,562,563,565,567,569,571,572,574,576,579,581,582,584,586,589,591,592,593,595,598,600,601,602,604,607,609,610,612,614,616,618,619,621,623,625,627,628,629],[144,144,144,473,146,144,146,146,144],85.1,[144,144,148,475,146,144,146,146,144],40.1,[144,148,155,477,146,144,144,146,144],0.052,[144,155,155,479,146,144,144,146,144],3.326,[144,184,155,481,146,144,144,146,144],3.378,[148,144,144,483,146,144,146,146,144],81.6,[148,144,148,485,146,144,146,146,144],37.2,[148,148,155,487,146,144,144,146,144],0.024,[148,155,155,489,146,144,144,146,144],3.088,[148,184,155,491,146,144,144,146,144],3.112,[155,144,144,493,146,144,146,146,144],88.6,[155,144,148,495,146,144,146,146,144],59.8,[155,148,155,497,146,144,144,146,144],60.248,[155,155,155,499,146,144,144,146,144],0.388,[155,184,155,501,146,144,144,146,144],60.636,[184,144,144,503,146,144,146,146,144],89,[184,144,148,495,146,144,146,146,144],[184,148,155,506,146,144,144,146,144],0.032,[184,155,155,508,146,144,144,146,144],5.12,[184,184,155,510,146,144,144,146,144],5.152,[197,144,144,512,146,144,146,146,144],89.3,[197,144,148,514,146,144,146,146,144],67.5,[197,148,155,516,146,144,144,146,144],0.115,[197,155,155,518,146,144,144,146,144],1.807,[197,184,155,520,146,144,144,146,144],1.922,[218,144,144,522,146,144,146,146,144],92.3,[218,144,148,524,146,144,146,146,144],75.4,[218,148,155,526,146,144,144,146,144],0.075,[218,155,155,528,146,144,144,146,144],1.558,[218,184,155,530,146,144,144,146,144],1.633,[270,144,144,532,146,144,146,146,144],92.2,[270,144,148,534,146,144,146,146,144],74.9,[270,148,155,536,146,144,144,146,144],0.06,[270,155,155,538,146,144,144,146,144],1.546,[270,184,155,540,146,144,144,146,144],1.606,[275,144,144,542,146,144,146,146,144],42.1,[275,144,148,544,146,144,146,146,144],3.9,[275,148,155,477,146,144,144,146,144],[275,155,155,547,146,144,144,146,144],0.008,[275,184,155,549,146,144,144,146,144],0.056,[91,144,144,551,146,144,146,146,144],37.4,[91,144,148,553,146,144,146,146,144],2.8,[91,148,155,487,146,144,144,146,144],[91,155,155,547,146,144,144,146,144],[91,184,155,506,146,144,144,146,144],[558,144,144,559,146,144,146,146,144],9,34,[558,144,148,561,146,144,146,146,144],2.5,[558,148,155,497,146,144,144,146,144],[558,155,155,564,146,144,144,146,144],0.006,[558,184,155,566,146,144,144,146,144],60.254,[415,144,144,568,146,144,146,146,144],50,[415,144,148,570,146,144,146,146,144],6.4,[415,148,155,506,146,144,144,146,144],[415,155,155,573,146,144,144,146,144],0.009,[415,184,155,575,146,144,144,146,144],0.041,[577,144,144,578,146,144,146,146,144],11,64.6,[577,144,148,580,146,144,146,146,144],21.6,[577,148,155,516,146,144,144,146,144],[577,155,155,583,146,144,144,146,144],0.003,[577,184,155,585,146,144,144,146,144],0.118,[587,144,144,588,146,144,146,146,144],12,86.7,[587,144,148,590,146,144,146,146,144],60.5,[587,148,155,526,146,144,144,146,144],[587,155,155,583,146,144,144,146,144],[587,184,155,594,146,144,144,146,144],0.078,[596,144,144,597,146,144,146,146,144],13,87.5,[596,144,148,599,146,144,146,146,144],61.4,[596,148,155,536,146,144,144,146,144],[596,155,155,583,146,144,144,146,144],[596,184,155,603,146,144,144,146,144],0.063,[605,144,144,606,146,144,146,146,144],14,87.6,[605,144,148,608,146,144,146,146,144],64.8,[605,148,155,516,146,144,144,146,144],[605,155,155,611,146,144,144,146,144],0.028,[605,184,155,613,146,144,144,146,144],0.143,[340,144,144,615,146,144,146,146,144],91.8,[340,144,148,617,146,144,146,146,144],74.5,[340,148,155,526,146,144,144,146,144],[340,155,155,620,146,144,144,146,144],0.013,[340,184,155,622,146,144,144,146,144],0.088,[624,144,144,615,146,144,146,146,144],16,[624,144,148,626,146,144,146,146,144],74.2,[624,148,155,536,146,144,144,146,144],[624,155,155,620,146,144,144,146,144],[624,184,155,630,146,144,144,146,144],0.073,[],[414],[634],"RTX 3090 GPU (Sec. 4.1); CPU not stated",[],[637],"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":639,"group":640,"sourceId":5,"sourceLabel":6,"table":98,"selfRows":558,"metrics":641,"seqs":645,"entrants":666,"cells":679,"outcomes":769,"locators":770,"hardware":771,"wordings":772,"notes":773},"choy2019fcgf-table-5","choy2019fcgf:Table 5",[642],{"label":643,"unit":644,"statistic":18,"alignment":40},"registration recall","fraction",[646,650,652,654,656,658,660,662,664],{"dataset":647,"sequence":648,"environment":649},"3DMatch registration set","Kitchen","indoor",{"dataset":647,"sequence":651,"environment":649},"Home 1",{"dataset":647,"sequence":653,"environment":649},"Home 2",{"dataset":647,"sequence":655,"environment":649},"Hotel 1",{"dataset":647,"sequence":657,"environment":649},"Hotel 2",{"dataset":647,"sequence":659,"environment":649},"Hotel 3",{"dataset":647,"sequence":661,"environment":649},"Study",{"dataset":647,"sequence":663,"environment":649},"Lab",{"dataset":647,"sequence":665,"environment":649},"Average",[667,669,671,673,675,677],{"name":668,"methodId":132,"linkable":133,"proposed":66,"self":66},"FPFH [23]",{"name":670,"methodId":61,"linkable":66,"proposed":66,"self":66},"USC [29]",{"name":672,"methodId":61,"linkable":66,"proposed":66,"self":66},"CGF [17]",{"name":674,"methodId":61,"linkable":66,"proposed":66,"self":66},"3DMatch [36]",{"name":676,"methodId":61,"linkable":66,"proposed":66,"self":66},"PPFNet [7]",{"name":678,"methodId":5,"linkable":133,"proposed":133,"self":133},"Ours (FCGF)",[680,682,684,686,688,689,691,693,694,696,698,700,702,704,706,708,710,712,713,715,716,717,719,720,722,724,725,726,728,730,731,732,734,736,738,740,742,744,745,746,748,750,752,753,755,756,758,760,761,762,764,765,766,768],[144,144,144,681,146,144,146,146,144],0.36,[144,144,148,683,146,144,146,146,144],0.56,[144,144,155,685,146,144,146,146,144],0.43,[144,144,184,687,146,144,146,146,144],0.29,[144,144,197,681,146,144,146,146,144],[144,144,218,690,146,144,146,146,144],0.61,[144,144,270,692,146,144,146,146,144],0.31,[144,144,275,692,146,144,146,146,144],[144,144,91,695,146,144,146,146,144],0.4,[148,144,144,697,146,144,146,146,144],0.52,[148,144,148,699,146,144,146,146,144],0.35,[148,144,155,701,146,144,146,146,144],0.47,[148,144,184,703,146,144,146,146,144],0.53,[148,144,197,705,146,144,146,146,144],0.2,[148,144,218,707,146,144,146,146,144],0.38,[148,144,270,709,146,144,146,146,144],0.46,[148,144,275,711,146,144,146,146,144],0.49,[148,144,91,685,146,144,146,146,144],[155,144,144,714,146,144,146,146,144],0.72,[155,144,148,268,146,144,146,146,144],[155,144,155,709,146,144,146,146,144],[155,144,184,718,146,144,146,146,144],0.55,[155,144,197,711,146,144,146,146,144],[155,144,218,721,146,144,146,146,144],0.65,[155,144,270,723,146,144,146,146,144],0.48,[155,144,275,280,146,144,146,146,144],[155,144,91,683,146,144,146,146,144],[184,144,144,727,146,144,146,146,144],0.85,[184,144,148,729,146,144,146,146,144],0.78,[184,144,155,690,146,144,146,146,144],[184,144,184,264,146,144,146,146,144],[184,144,197,733,146,144,146,146,144],0.59,[184,144,218,735,146,144,146,146,144],0.58,[184,144,270,737,146,144,146,146,144],0.63,[184,144,275,739,146,144,146,146,144],0.51,[184,144,91,741,146,144,146,146,144],0.67,[197,144,144,743,146,144,146,146,144],0.9,[197,144,148,735,146,144,146,146,144],[197,144,155,373,146,144,146,146,144],[197,144,184,747,146,144,146,146,144],0.75,[197,144,197,749,146,144,146,146,144],0.68,[197,144,218,751,146,144,146,146,144],0.88,[197,144,270,749,146,144,146,146,144],[197,144,275,754,146,144,146,146,144],0.62,[197,144,91,195,146,144,146,146,144],[218,144,144,757,146,144,146,146,144],0.93,[218,144,148,759,146,144,146,146,144],0.91,[218,144,155,195,146,144,146,146,144],[218,144,184,759,146,144,146,146,144],[218,144,197,763,146,144,146,146,144],0.87,[218,144,218,268,146,144,146,146,144],[218,144,270,747,146,144,146,146,144],[218,144,275,767,146,144,146,146,144],0.8,[218,144,91,276,146,144,146,146,144],[],[98],[],[],[774],"Registration recall on the 3DMatch registration set; RANSAC with early termination; pair correct if overlap >= 30% and RMSE \u003C 0.2 m",[776,782,789,794],{"group":777,"slug":778,"sourceLabel":6,"table":779,"selfRows":184,"datasets":780},"choy2019fcgf:Table 1","choy2019fcgf-table-1","Table 1",[428,781],"3DMatch with rotation augmentation",{"group":783,"slug":784,"sourceLabel":785,"table":786,"selfRows":184,"datasets":787},"lim2025kissmatcher:Table I","lim2025kissmatcher-table-i","Lim et al., 2025","Table I",[788],"KITTI",{"group":790,"slug":791,"sourceLabel":413,"table":792,"selfRows":184,"datasets":793},"qin2023geotransformer:Table 3","qin2023geotransformer-table-3","Table 3",[86],{"group":795,"slug":796,"sourceLabel":6,"table":797,"selfRows":155,"datasets":798},"choy2019fcgf:Text Sec. 6.7","choy2019fcgf-text-sec-6-7","Text Sec. 6.7",[428],1790510664966]