[{"data":1,"prerenderedAt":643},["ShallowReactive",2],{"method-badslam2019":3},{"method":4,"reference":58,"equipment":80,"figures":131,"results":132},{"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":24,"limitations":29,"sensors":36,"platform":38,"estimator":42,"association":43,"timeModel":44,"deskew":45,"loopClosure":46,"globalOptimization":47,"mapRepresentation":48,"prior":49,"outputGeometry":50,"compute":51,"codeUrl":52,"codeLicense":53,"relatedVersions":54},"badslam2019","Schöps et al., 2019","BAD SLAM","BAD SLAM: Bundle Adjusted Direct RGB-D SLAM",2019,"recent","C08","full_slam_with_global_correction","BAD SLAM 提出可即時執行的直接式 BA，以面元表示地圖，同時使用深度的幾何約束與影像梯度的光度約束，並交替最佳化地圖與相機位姿。作者另建立以同步全域快門（global shutter）RGB 與深度相機錄製、經精確校正的 ETH3D SLAM 基準，指出直接式 RGB-D SLAM 對捲簾快門、RGB 與深度不同步及校正誤差高度敏感。在此基準上方法排名與既有資料集不同，顯示資料集設定本身會左右比較結論。","BAD SLAM performs real-time direct bundle adjustment over surfels and keyframes and introduces a well-calibrated, synchronised global-shutter RGB-D benchmark, showing direct RGB-D SLAM is highly sensitive to rolling shutter, RGB-depth synchronisation and calibration errors.","full_text_reviewed","peer_reviewed_published","main_body","論文未報告營建測試。其結論指出感測器同步、快門型式與校正誤差會改變方法排名，直接呼應本文對工程點雲誤差來源與公平比較的要求（推論連結）。作者另指出白牆這類弱紋理且幾何變化少的區域會讓面元在表面內任意滑動，因此限制面元只沿法向移動；困難序列的失敗原因包含無紋理且結構模糊的場景（Sec. 3.1、Sec. 6），與室內裝修前的素牆環境相關（推論）。",[20,21,22,23],"public_benchmark","controlled_experiment","simulation","independent_reference",[25,26,27,28],"Direct BA with rich information yields very accurate trajectories on the new benchmark (conclusion; Figs. 6 and 7)","Open benchmark with hidden test set and online evaluation (abstract)","Lowest average and median ATE RMSE of the five systems on all four synthetic distortion variants (Table 3)","Rankings on the benchmark do not depend on using ATE or relative translation and rotation metrics (Supp. Sec. 2.1)",[30,31,32,33,34,35],"Direct RGB-D SLAM is highly sensitive to rolling shutter, RGB-depth synchronisation and calibration errors (abstract; Table 3)","Benchmark contains hard sequences that remain open challenges for visual-only RGB-D SLAM (conclusion)","On TUM RGB-D ORB-SLAM2 outperforms all direct methods including BAD SLAM, and without intrinsics and depth-deformation optimisation BAD SLAM is clearly worse (Table 2)","Direct alignment has a small convergence region; a pose-graph step after loop detection could in principle push old keyframes out of it (Sec. 4)","Failures come from textureless scenes with ambiguous structure, fast motion and moving objects; an IMU or wider field of view would be needed (Sec. 6)","Scalability not evaluated; real-time BA for longer sequences would need e.g. windowed BA (Sec. 6, Sec. 7)",[37],"RGB-D",[39,40,41],"custom multi-camera rig with up to eight synchronized global-shutter cameras (two colour and two infrared cameras form the RGB-D sensor; four more used only for SfM ground truth) and an Asus Xtion Live Pro used solely as infrared pattern emitter, tracked by a Vicon system; carrying mode not stated in paper or supplement","public TUM RGB-D sequences (Kinect v1)","synthetic renders of TUM RGB-D reconstructions","Alternating direct BA (Alg. 1): Gauss-Newton on a cost of Tukey-weighted point-to-plane residuals and Huber-weighted photometric gradient residuals (photometric weight 1e-2), each normalised by a stereo depth-noise model or an empirical sigma; surfels move only along their normals (joint 2x2 solve of offset and descriptor per surfel); independent per-keyframe SE(3) pose updates; optional intrinsics and per-pixel depth-deformation calibration solved with the Schur complement; interleaved discrete surfel creation, merging, outlier deletion and radius update. A PCG Gauss-Newton solver was slightly worse in the ablation.","Surfel to pixel correspondences by projecting surfel centres into every keyframe, kept only if the pixel has depth, the Tukey weight of the geometric residual is positive and normals agree; photometric residual compares the surfel descriptor with the gradient magnitude sampled at the surfel centre and two disc boundary points. Front-end odometry: direct photometric and geometric SE(3) alignment of each frame to the last keyframe using intensity gradients.","discrete poses; every 10th frame becomes a keyframe; benchmark colour and depth are captured at the same instants, so no temporal interpolation between them is needed","not_applicable (rolling shutter avoided by global-shutter, synchronised benchmark cameras rather than modelled)","bag-of-words detection with binary features; relative pose from keypoint matches refined by direct alignment and checked for consistency against neighbouring keyframes m-1 and m+1; then pose-graph optimisation (Sec. 4; Fig. 2)","pose-graph optimisation on loop detection, plus direct BA back-end over keyframe poses, surfels and camera intrinsics (Fig. 2; Sec. 3-4)","surfels (oriented discs with centre, normal, radius and a scalar gradient descriptor) created on 4x4 pixel cells of each keyframe, plus keyframes storing raw RGB-D; about 335,000 surfels for the Fig. 1 scene","none","surfel map and keyframe trajectory","Intel Core i7 6700K with MSI GeForce GTX 1080 Gaming X 8G; BA on GPU with CUDA 8.0; with about 27 Hz input and one keyframe per 10 frames, 370 ms of BA time per keyframe are available and further BA iterations are skipped in real-time mode; odometry time negligible; memory grows linearly with keyframe count","https:\u002F\u002Fgithub.com\u002FETH3D\u002Fbadslam","BSD-style (LICENSE header, ETH Zurich 2019)",[55],{"relation":56,"title":57,"doi_or_url":52},"code_release","badslam",{"id":5,"kind":59,"shortName":7,"title":8,"authors":60,"year":9,"venue":64,"venueType":65,"publisher":66,"volumeIssuePages":67,"doi":68,"arxivId":69,"url":70,"firstPublicDate":71,"publicationStatus":16,"metadataStatus":72,"fulltextStatus":15,"era":10,"classicReason":73,"codeUrl":52,"cluster":11,"topics":74,"mdpi":76,"verification":77,"label":6,"fulltextRoute":78,"versionRead":79,"addedByCensus":76},"method",[61,62,63],"Thomas Schöps","Torsten Sattler","Marc Pollefeys","2019 IEEE\u002FCVF Conference on Computer Vision and Pattern Recognition (CVPR)","conference","IEEE","pp. 134-144","10.1109\u002Fcvpr.2019.00022",null,"https:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_CVPR_2019\u002Fpapers\u002FSchops_BAD_SLAM_Bundle_Adjusted_Direct_RGB-D_SLAM_CVPR_2019_paper.pdf","2019-06","metadata_verified","not_applicable",[11,75],"C10",false,"corrected","other","CVF open-access version of the CVPR 2019 paper (pp. 134-144; identical to the accepted version except for the watermark per CVF statement) plus the CVF supplementary document (10 pages, file 2315-supp.pdf); IEEE Xplore version of record not compared",[81,88,92,98,103,108,112,117,123,126],{"category":82,"model":83,"canonical":83,"role":84,"dataset":85,"specs":86,"locator":87},"stereo_camera","two synchronized global-shutter colour cameras (front-facing stereo pair) on a custom multi-camera rig similar to Gohl et al. [22]","dataset sensor","ETH3D SLAM benchmark (this paper)","one camera supplies the RGB image of the RGB-D frames; both usable for stereo SLAM; raw Bayer images debayered, flat-field corrected, no white balancing","Sec. 5; Supp. Sec. 3.1, 4.1, Fig. 2",{"category":82,"model":89,"canonical":89,"role":84,"dataset":85,"specs":90,"locator":91},"two synchronized global-shutter infrared cameras (stereo pair below the colour cameras) on the same rig","active stereo depth by PatchMatch stereo with ZNCC cost and 11x11 window, reprojected to the colour camera; colour and depth recorded at exactly the same time","Sec. 5; Supp. Sec. 3.1, 4.6",{"category":93,"model":94,"canonical":94,"role":95,"dataset":85,"specs":96,"locator":97},"camera","four additional synchronized cameras on the rig","reference or ground truth","used only to localise training sequences whose ground truth comes from Structure-from-Motion","Sec. 5; Supp. Sec. 3.1",{"category":99,"model":100,"canonical":100,"role":84,"dataset":85,"specs":101,"locator":102},"rgbd","Asus Xtion Live Pro","mounted on the rig and used only as infrared pattern emitter for active stereo; its own depth estimation not used","Sec. 5; Supp. Sec. 3.1, footnote 1",{"category":82,"model":104,"canonical":104,"role":105,"dataset":69,"specs":106,"locator":107},"Intel D435","compared device","infrared emitter considered but not used because its projected pattern has lower resolution than the Xtion pattern","Supp. Sec. 3.1, Fig. 3",{"category":78,"model":109,"canonical":109,"role":95,"dataset":85,"specs":110,"locator":111},"Vicon motion capturing system","tracks passive markers on the rig; outliers removed manually and trajectory smoothed with an SE(3) cubic B-spline (20 knots per second); camera-Vicon calibration ATE RMSE required to be at most 1 mm","Sec. 5; Supp. Sec. 3, 4.4, 4.5, 4.7",{"category":113,"model":114,"canonical":114,"role":84,"dataset":85,"specs":115,"locator":116},"imu","IMU (model not stated in paper or supplement)","benchmark stated to provide full IMU data","Sec. 2, Table 1",{"category":118,"model":119,"canonical":119,"role":120,"dataset":69,"specs":121,"locator":122},"compute","Intel Core i7 6700K","compute for runtime","not_reported","Sec. 6 Test environment",{"category":118,"model":124,"canonical":124,"role":120,"dataset":69,"specs":125,"locator":122},"MSI Geforce GTX 1080 Gaming X 8G","BA implemented on the GPU with CUDA 8.0",{"category":99,"model":127,"canonical":127,"role":84,"dataset":128,"specs":129,"locator":130},"Kinect v1","TUM RGB-D","sensor of the TUM RGB-D benchmark; rolling shutter and unsynchronised depth and colour; its shutter times (about 30.5 ms depth, 26.1 ms colour, from [52]) used to render synthetic rolling-shutter variants; visible depth distortion pattern in BAD SLAM self-calibration","Sec. 5 Impact of distortions; Supp. Fig. 9",[],{"totalRows":133,"groupCount":134,"groups":135,"others":603},46,11,[136,266,365,487],{"slug":137,"group":138,"sourceId":5,"sourceLabel":6,"table":139,"selfRows":140,"metrics":141,"seqs":149,"entrants":160,"cells":175,"outcomes":260,"locators":261,"hardware":262,"wordings":263,"notes":264},"badslam2019-table-3","badslam2019:Table 3","Table 3",8,[142,146],{"label":143,"unit":144,"statistic":145,"alignment":121},"ATE RMSE [cm], average over seven synthetic datasets","cm","mean",{"label":147,"unit":144,"statistic":148,"alignment":121},"ATE RMSE [cm], median over seven synthetic datasets","median",[150,154,156,158],{"dataset":151,"sequence":152,"environment":153},"synthetic TUM RGB-D renders (7 datasets per category)","clean","synthetic renders of dense TUM RGB-D scene reconstructions along the original trajectories (Sec. 5)",{"dataset":151,"sequence":155,"environment":153},"async",{"dataset":151,"sequence":157,"environment":153},"rs",{"dataset":151,"sequence":159,"environment":153},"async & rs",[161,165,167,170,173],{"name":162,"methodId":163,"linkable":164,"proposed":76,"self":76},"BundleFusion","bundlefusion2017",true,{"name":166,"methodId":69,"linkable":76,"proposed":76,"self":76},"DVO SLAM",{"name":168,"methodId":169,"linkable":164,"proposed":76,"self":76},"ElasticFusion","elasticfusion2015",{"name":171,"methodId":172,"linkable":164,"proposed":76,"self":76},"ORB-SLAM2","orbslam2_2017",{"name":174,"methodId":5,"linkable":164,"proposed":164,"self":164},"BAD SLAM (Ours)",[176,180,183,186,189,192,194,196,198,200,202,204,206,208,210,212,214,216,218,219,221,222,224,226,228,230,232,234,236,238,240,242,244,246,248,250,252,254,256,258],[177,177,177,178,179,177,179,179,177],0,0.34,-1,[181,177,177,182,179,177,179,179,177],1,0.32,[184,177,177,185,179,177,179,179,177],2,1.11,[187,177,177,188,179,177,179,179,177],3,0.47,[190,177,177,191,179,177,179,179,177],4,0.15,[177,181,177,193,179,177,179,179,177],0.22,[181,181,177,195,179,177,179,179,177],0.23,[184,181,177,197,179,177,179,179,177],0.9,[187,181,177,199,179,177,179,179,177],0.3,[190,181,177,201,179,177,179,179,177],0.02,[177,177,181,203,179,177,179,179,177],1.1,[181,177,181,205,179,177,179,179,177],2.33,[184,177,181,207,179,177,179,179,177],1.98,[187,177,181,209,179,177,179,179,177],0.6,[190,177,181,211,179,177,179,179,177],0.4,[177,181,181,213,179,177,179,179,177],1.14,[181,181,181,215,179,177,179,179,177],0.72,[184,181,181,217,179,177,179,179,177],1.17,[187,181,181,211,179,177,179,179,177],[190,181,181,220,179,177,179,179,177],0.21,[177,177,184,203,179,177,179,179,177],[181,177,184,223,179,177,179,179,177],5.1,[184,177,184,225,179,177,179,179,177],2.7,[187,177,184,227,179,177,179,179,177],3.25,[190,177,184,229,179,177,179,179,177],0.99,[177,181,184,231,179,177,179,179,177],1.02,[181,181,184,233,179,177,179,179,177],1.37,[184,181,184,235,179,177,179,179,177],1.77,[187,181,184,237,179,177,179,179,177],1.57,[190,181,184,239,179,177,179,179,177],0.87,[177,177,187,241,179,177,179,179,177],1.48,[181,177,187,243,179,177,179,179,177],4.94,[184,177,187,245,179,177,179,179,177],3.19,[187,177,187,247,179,177,179,179,177],3.49,[190,177,187,249,179,177,179,179,177],1.01,[177,181,187,251,179,177,179,179,177],1.4,[181,181,187,253,179,177,179,179,177],1.39,[184,181,187,255,179,177,179,179,177],2.52,[187,181,187,257,179,177,179,179,177],1.55,[190,181,187,259,179,177,179,179,177],0.98,[],[139],[],[],[265],"Synthetic renders of dense TUM RGB-D reconstructions along the original trajectories; each value aggregates ATE RMSE over seven synthetic datasets per category (avg. or med.); rs uses Kinect v1 shutter times (about 30.5 ms depth, 26.1 ms colour); async renders colour midway between depth frames",{"slug":267,"group":268,"sourceId":5,"sourceLabel":6,"table":269,"selfRows":270,"metrics":271,"seqs":275,"entrants":283,"cells":303,"outcomes":358,"locators":360,"hardware":361,"wordings":362,"notes":363},"badslam2019-table-2","badslam2019:Table 2","Table 2",6,[272],{"label":273,"unit":144,"statistic":274,"alignment":121},"ATE RMSE [cm]","RMSE",[276,279,281],{"dataset":128,"sequence":277,"environment":278},"fr1\u002Fdesk","TUM RGB-D real-world sequences recorded with a Kinect v1 (rolling shutter; depth and colour streams not synchronised, Sec. 5); scene type and carrying mode not described in this paper",{"dataset":128,"sequence":280,"environment":278},"fr2\u002Fxyz",{"dataset":128,"sequence":282,"environment":278},"fr3\u002Foffice",[284,285,286,287,290,292,293,295,297,300,302],{"name":162,"methodId":163,"linkable":164,"proposed":76,"self":76},{"name":166,"methodId":69,"linkable":76,"proposed":76,"self":76},{"name":168,"methodId":169,"linkable":164,"proposed":76,"self":76},{"name":288,"methodId":289,"linkable":164,"proposed":76,"self":76},"Kintinuous","kintinuous2015",{"name":291,"methodId":69,"linkable":76,"proposed":76,"self":76},"MRSMap",{"name":171,"methodId":172,"linkable":164,"proposed":76,"self":76},{"name":294,"methodId":69,"linkable":76,"proposed":76,"self":76},"PSM SLAM",{"name":296,"methodId":69,"linkable":76,"proposed":76,"self":76},"RGB-D SLAM",{"name":298,"methodId":299,"linkable":164,"proposed":76,"self":76},"VoxelHashing","voxelhashing2013",{"name":301,"methodId":5,"linkable":164,"proposed":76,"self":164},"BAD SLAM ablation: Ours (fixed intr.)",{"name":174,"methodId":5,"linkable":164,"proposed":164,"self":164},[304,306,307,309,311,313,315,316,317,319,321,323,324,326,327,329,331,332,333,334,335,337,340,342,344,345,346,347,350,352,354,356,357],[177,177,177,305,179,177,179,179,177],1.6,[177,177,181,203,179,177,179,179,177],[177,177,184,308,179,177,179,179,177],2.2,[181,177,177,310,179,177,179,179,177],2.1,[181,177,181,312,179,177,179,179,177],1.8,[181,177,184,314,179,177,179,179,177],3.5,[184,177,177,184,179,177,179,179,177],[184,177,181,203,179,177,179,179,177],[184,177,184,318,179,177,179,179,177],1.7,[187,177,177,320,179,177,179,179,177],3.7,[187,177,181,322,179,177,179,179,177],2.9,[187,177,184,187,179,177,179,179,177],[190,177,177,325,179,177,179,179,177],4.3,[190,177,181,184,179,177,179,179,177],[190,177,184,328,179,177,179,179,177],4.2,[330,177,177,305,179,177,179,179,177],5,[330,177,181,211,179,177,179,179,177],[330,177,184,181,179,177,179,179,177],[270,177,177,305,179,177,179,179,177],[270,177,181,69,177,177,179,179,177],[270,177,184,336,179,177,179,179,177],3.1,[338,177,177,339,179,177,179,179,177],7,2.3,[338,177,181,341,179,177,179,179,177],0.8,[338,177,184,343,179,177,179,179,177],3.2,[140,177,177,339,179,177,179,179,177],[140,177,181,308,179,177,179,179,177],[140,177,184,339,179,177,179,179,177],[348,177,177,349,179,177,179,179,177],9,3.6,[348,177,181,351,179,177,179,179,177],1.2,[348,177,184,353,179,177,179,179,177],2.5,[355,177,177,318,179,177,179,179,177],10,[355,177,181,203,179,177,179,179,177],[355,177,184,318,179,177,179,179,177],[359],"no value in source ('-')",[269],[],[],[364],"TUM RGB-D ATE RMSE in cm (rank column omitted); values of other methods copied by the authors from BundleFusion, PSM SLAM and ORB-SLAM2 papers; 'fixed intr.' disables intrinsics and depth-deformation optimisation",{"slug":366,"group":367,"sourceId":368,"sourceLabel":369,"table":139,"selfRows":270,"metrics":370,"seqs":372,"entrants":384,"cells":402,"outcomes":478,"locators":480,"hardware":481,"wordings":482,"notes":485},"pointslam2023-table-3","pointslam2023:Table 3","pointslam2023","Sandström et al., 2023",[371],{"label":273,"unit":144,"statistic":274,"alignment":121},[373,376,378,380,381,382],{"dataset":374,"sequence":277,"environment":375},"TUM-RGBD","indoor office (real RGB-D)",{"dataset":374,"sequence":377,"environment":375},"fr1\u002Fdesk2",{"dataset":374,"sequence":379,"environment":375},"fr1\u002Froom",{"dataset":374,"sequence":280,"environment":375},{"dataset":374,"sequence":282,"environment":375},{"dataset":374,"sequence":383,"environment":375},"Avg.",[385,387,390,392,394,396,398,400],{"name":386,"methodId":69,"linkable":76,"proposed":76,"self":76},"DI-Fusion [19]",{"name":388,"methodId":389,"linkable":164,"proposed":76,"self":76},"NICE-SLAM [81]","niceslam2022",{"name":391,"methodId":69,"linkable":76,"proposed":76,"self":76},"Vox-Fusion* [71] (re-run)",{"name":393,"methodId":368,"linkable":164,"proposed":164,"self":76},"Point-SLAM (ours)",{"name":395,"methodId":5,"linkable":164,"proposed":76,"self":164},"BAD-SLAM [50]",{"name":397,"methodId":289,"linkable":164,"proposed":76,"self":76},"Kintinuous [68]",{"name":399,"methodId":172,"linkable":164,"proposed":76,"self":76},"ORB-SLAM2 [35]",{"name":401,"methodId":169,"linkable":164,"proposed":76,"self":76},"ElasticFusion [67]",[403,405,406,407,408,410,411,413,415,417,419,421,423,425,426,428,430,432,434,436,438,440,442,444,446,447,448,449,450,451,452,453,455,457,458,459,461,462,463,465,466,467,468,470,472,474,475,476],[177,177,177,404,179,177,179,179,177],4.4,[177,177,181,69,177,177,179,179,177],[177,177,184,69,177,177,179,179,177],[177,177,187,184,179,177,179,179,177],[177,177,190,409,179,177,179,179,177],5.8,[177,177,330,69,177,177,179,179,177],[181,177,177,412,179,177,179,179,177],4.26,[181,177,181,414,179,177,179,179,177],4.99,[181,177,184,416,179,177,179,179,177],34.49,[181,177,187,418,179,177,179,177,177],31.73,[181,177,190,420,179,177,179,179,177],3.87,[181,177,330,422,179,177,179,181,177],15.87,[184,177,177,424,179,177,179,179,177],3.52,[184,177,181,270,179,177,179,179,177],[184,177,184,427,179,177,179,179,177],19.53,[184,177,187,429,179,177,179,179,177],1.49,[184,177,190,431,179,177,179,179,177],26.01,[184,177,330,433,179,177,179,179,177],11.31,[187,177,177,435,179,177,179,179,177],4.34,[187,177,181,437,179,177,179,179,177],4.54,[187,177,184,439,179,177,179,179,177],30.92,[187,177,187,441,179,177,179,179,177],1.31,[187,177,190,443,179,177,179,179,177],3.48,[187,177,330,445,179,177,179,179,177],8.92,[190,177,177,318,179,177,179,179,177],[190,177,181,69,177,177,179,179,177],[190,177,184,69,177,177,179,179,177],[190,177,187,203,179,177,179,179,177],[190,177,190,318,179,177,179,179,177],[190,177,330,69,177,177,179,179,177],[330,177,177,320,179,177,179,179,177],[330,177,181,454,179,177,179,179,177],7.1,[330,177,184,456,179,177,179,179,177],7.5,[330,177,187,322,179,177,179,179,177],[330,177,190,187,179,177,179,179,177],[330,177,330,460,179,177,179,179,177],4.84,[270,177,177,305,179,177,179,179,177],[270,177,181,308,179,177,179,179,177],[270,177,184,464,179,177,179,179,177],4.7,[270,177,187,211,179,177,179,179,177],[270,177,190,181,179,177,179,179,177],[270,177,330,207,179,177,179,179,177],[338,177,177,469,179,177,179,179,177],2.53,[338,177,181,471,179,177,179,179,177],6.83,[338,177,184,473,179,177,179,179,177],21.49,[338,177,187,217,179,177,179,179,177],[338,177,190,255,179,177,179,179,177],[338,177,330,477,179,177,179,179,177],6.91,[479],"not_reported (N\u002FA)",[139],[],[483,484],"ATE RMSE [cm], as written 31.73 (6.19 over successful runs)","ATE RMSE [cm], as written 15.87 (10.76 over successful runs)",[486],"TUM-RGBD tracking, ATE RMSE in cm; values in parentheses are averages over successful runs only; N\u002FA not available; ground truth from an external motion capture system",{"slug":488,"group":489,"sourceId":490,"sourceLabel":491,"table":269,"selfRows":190,"metrics":492,"seqs":495,"entrants":505,"cells":533,"outcomes":597,"locators":598,"hardware":599,"wordings":600,"notes":601},"gsslam2024-table-2","gsslam2024:Table 2","gsslam2024","Yan et al., 2024",[493],{"label":494,"unit":144,"statistic":274,"alignment":121},"ATE [cm] (text calls it ATE RSME)",[496,499,501,503],{"dataset":128,"sequence":497,"environment":498},"fr1_desk","real indoor (office, desk)",{"dataset":128,"sequence":500,"environment":498},"fr2_xyz",{"dataset":128,"sequence":502,"environment":498},"fr3_off",{"dataset":128,"sequence":504,"environment":498},"average of 3 sequences",[506,508,510,512,514,516,519,521,523,526,529,531],{"name":507,"methodId":69,"linkable":76,"proposed":76,"self":76},"DI-Fusion [ 9 ]",{"name":509,"methodId":169,"linkable":164,"proposed":76,"self":76},"ElasticFusion [ 46 ]",{"name":511,"methodId":5,"linkable":164,"proposed":76,"self":164},"BAD-SLAM [ 30 ]",{"name":513,"methodId":289,"linkable":164,"proposed":76,"self":76},"Kintinuous [ 45 ]",{"name":515,"methodId":172,"linkable":164,"proposed":76,"self":76},"ORB-SLAM2 [ 20 ]",{"name":517,"methodId":518,"linkable":164,"proposed":76,"self":76},"iMAP ∗ [ 35 ]","imap2021",{"name":520,"methodId":389,"linkable":164,"proposed":76,"self":76},"NICE-SLAM [ 55 ]",{"name":522,"methodId":69,"linkable":76,"proposed":76,"self":76},"Vox-Fusion ∗ [ 48 ]",{"name":524,"methodId":525,"linkable":164,"proposed":76,"self":76},"CoSLAM [ 41 ]","coslam2023",{"name":527,"methodId":528,"linkable":164,"proposed":76,"self":76},"ESLAM [ 11 ]","eslam2023",{"name":530,"methodId":368,"linkable":164,"proposed":76,"self":76},"Point-SLAM",{"name":532,"methodId":69,"linkable":76,"proposed":164,"self":76},"Ours",[534,535,536,537,539,540,541,542,543,544,545,546,548,549,550,551,552,553,554,555,556,558,559,560,562,563,565,567,569,570,571,573,575,576,578,580,582,583,584,585,586,587,589,590,591,593,594,596],[177,177,177,404,179,177,179,179,177],[177,177,181,184,179,177,179,179,177],[177,177,184,409,179,177,179,179,177],[177,177,187,538,179,177,179,179,177],4.1,[181,177,177,353,179,177,179,179,177],[181,177,181,351,179,177,179,179,177],[181,177,184,353,179,177,179,179,177],[181,177,187,310,179,177,179,179,177],[184,177,177,318,179,177,179,179,177],[184,177,181,203,179,177,179,179,177],[184,177,184,318,179,177,179,179,177],[184,177,187,547,179,177,179,179,177],1.5,[187,177,177,320,179,177,179,179,177],[187,177,181,322,179,177,179,179,177],[187,177,184,187,179,177,179,179,177],[187,177,187,343,179,177,179,179,177],[190,177,177,305,179,177,179,179,177],[190,177,181,211,179,177,179,179,177],[190,177,184,181,179,177,179,179,177],[190,177,187,181,179,177,179,179,177],[330,177,177,557,179,177,179,179,177],7.2,[330,177,181,310,179,177,179,179,177],[330,177,184,348,179,177,179,179,177],[330,177,187,561,179,177,179,179,177],6.1,[270,177,177,325,179,177,179,179,177],[270,177,181,564,179,177,179,179,177],31.7,[270,177,184,566,179,177,179,179,177],3.9,[270,177,187,568,179,177,179,179,177],13.3,[338,177,177,314,179,177,179,179,177],[338,177,181,547,179,177,179,179,177],[338,177,184,572,179,177,179,179,177],26,[338,177,187,574,179,177,179,179,177],10.3,[140,177,177,225,179,177,179,179,177],[140,177,181,577,179,177,179,179,177],1.9,[140,177,184,579,179,177,179,179,177],2.6,[140,177,187,581,179,177,179,179,177],2.4,[348,177,177,339,179,177,179,179,177],[348,177,181,203,179,177,179,179,177],[348,177,184,581,179,177,179,179,177],[348,177,187,184,179,177,179,179,177],[355,177,177,579,179,177,179,179,177],[355,177,181,588,179,177,179,179,177],1.3,[355,177,184,343,179,177,179,179,177],[355,177,187,581,179,177,179,179,177],[134,177,177,592,179,177,179,179,177],3.3,[134,177,181,588,179,177,179,179,177],[134,177,184,595,179,177,179,179,177],6.6,[134,177,187,320,179,177,179,179,177],[],[269],[],[],[602],"TUM RGB-D ATE on three sequences; * = reproduced with official code",[604,610,615,621,626,631,636],{"group":605,"slug":606,"sourceLabel":607,"table":608,"selfRows":190,"datasets":609},"monogs2024:Table 1","monogs2024-table-1","Matsuki et al., 2024","Table 1",[128],{"group":611,"slug":612,"sourceLabel":613,"table":269,"selfRows":190,"datasets":614},"rtgslam2024:Table 2","rtgslam2024-table-2","Peng et al., 2024",[128],{"group":616,"slug":617,"sourceLabel":618,"table":619,"selfRows":187,"datasets":620},"coslam2023:Table 4","coslam2023-table-4","Wang et al., 2023a","Table 4",[128],{"group":622,"slug":623,"sourceLabel":624,"table":139,"selfRows":187,"datasets":625},"imap2021:Table 3","imap2021-table-3","Sucar et al., 2021",[128],{"group":627,"slug":628,"sourceLabel":629,"table":269,"selfRows":187,"datasets":630},"loopyslam2024:Table 2","loopyslam2024-table-2","Liso et al., 2024",[374],{"group":632,"slug":633,"sourceLabel":634,"table":269,"selfRows":187,"datasets":635},"niceslam2022:Table 2","niceslam2022-table-2","Zhu et al., 2022a",[128],{"group":637,"slug":638,"sourceLabel":639,"table":640,"selfRows":184,"datasets":641},"droidslam2021:Fig. 4 table","droidslam2021-fig-4-table","Teed & Deng, 2021","Fig. 4 table",[642],"ETH3D SLAM benchmark",1790510663852]