[{"data":1,"prerenderedAt":550},["ShallowReactive",2],{"method-deng2026_mcgs_slam":3},{"method":4,"reference":51,"equipment":70,"figures":92,"results":93},{"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":34,"platform":37,"estimator":38,"association":39,"timeModel":40,"deskew":41,"loopClosure":42,"globalOptimization":43,"mapRepresentation":44,"prior":45,"outputGeometry":46,"compute":47,"codeUrl":48,"codeLicense":49,"relatedVersions":50},"deng2026_mcgs_slam","Deng & Gan, 2026","MCGS SLAM","Motion-prior and Confidence-aware Gaussian Splatting (MCGS) SLAM for 3D scene reconstruction of indoor built environments",2026,"recent","C09","odometry_with_local_mapping","MCGS-SLAM 以 MonoGS 高斯潑濺 SLAM 為基礎，加入由位姿歷史或加速度計推得的運動先驗（含自適應權重與快速運動偵測）、依位置穩定性、形狀、不透明度、空間與邊緣重要性計算的高斯信心度、依速度、旋轉、位移、可見重疊與視覺豐富度的自適應關鍵影格選擇，以及信心加權的多任務最佳化。評估只用 TUM RGB-D、Replica 與 EuRoC 公開資料集；TUM fr1_desk 的 ATE RMSE 由 2.13 cm 降至 1.52 cm，但 Replica 七個場景有三個變差，EuRoC MH03 至 MH05 誤差仍達 1.79 至 3.78 m，且未在實際建物或工地測試。","MonoGS-based RGB-D 3DGS SLAM with a motion prior, per-Gaussian confidence and adaptive keyframing; evaluated only on TUM, Replica and EuRoC benchmarks, not on building data.","full_text_reviewed","peer_reviewed_published","supplementary","題名與動機指向室內建成環境與 Scan-to-BIM，但實驗只用 TUM RGB-D、Replica 合成資料與 EuRoC 大型室內序列，沒有建物或工地資料；作者在結論承認評估僅限基準資料集。可作為 3DGS SLAM 在弱紋理與快速運動情境的方法參考，不能作為營建場域精度的證據。",[20,21],"simulation","public_benchmark",[23,24,25,26],"TUM fr1_desk ATE RMSE 1.52 cm versus 2.13 cm for baseline MonoGS (Table 1)","TUM fr3_walking_xyz ATE RMSE 0.25 m versus 0.73 m for MonoGS (Table 2)","Replica average ATE improvement 11.03% (up to 83.09% in room1); PSNR +2.03%, SSIM +0.63%, LPIPS 11.76% (Tables 3 and 6)","With 25% of frames on fr1_desk, ATE RMSE 2.67 cm versus 6.64 cm for MonoGS (Table 7)",[28,29,30,31,32,33],"Evaluation limited to benchmark datasets; authors plan to collect a new dataset (Sec. 5)","Replica ATE worse than MonoGS in 3 of 7 scenes (office2, office3, room0) (Table 3)","EuRoC MH03 to MH05 ATE 1.79 to 3.78 m, far above ORB-SLAM3 stereo (0.02 to 0.09 m); MH05 worse than MonoGS (3.78 versus 3.19 m) (Table 11)","Full system is slower and heavier than MonoGS on fr1_desk (1.08 versus 1.98 FPS, 4.63 versus 3.21 GB) and its ATE (1.52 cm) is worse than the motion-prior-only (1.46 cm) or confidence-only (1.43 cm) variants (Table 10)","Much less accurate than dynamic-scene methods such as DynaSLAM (0.02 m) on fr3_walking_xyz (Table 2)","Many thresholds and weights are configuration parameters that the authors say can be tuned per dataset or sensor setup (Sec. 3.2 to 3.4; Sec. 5)",[35,36],"RGB-D camera (synchronized RGB-D input)","optional accelerometer and gyroscope for the motion prior; a motion-only mode infers velocity and acceleration from pose history when inertial data are absent",[],"frame-to-model tracking by minimizing photometric and geometric rendering losses on a 3DGS map (MonoGS baseline) plus an adaptively weighted motion-prior loss (translation, rotation, velocity and acceleration smoothness), with fast-motion detection that raises the prior weight and keyframe rate","dense direct photometric and geometric residuals over visible Gaussians, weighted by per-Gaussian confidence (position stability, shape quality, opacity, spatial importance, edge importance); keyframes chosen by velocity, rotation, displacement, visibility overlap and feature richness","discrete per-frame poses with a velocity and acceleration motion prior from a configurable pose-history window (or IMU integration when available)","not_applicable (camera frames; no scanning sensor)","none (authors state the confidence-aware map maintenance avoids descriptor-based loop closure; EuRoC table lists MCGS-SLAM without loop closure)","confidence-weighted bundle adjustment over keyframes in a sliding window during map maintenance; no global pose graph","3D Gaussian splatting map (MonoGS baseline) with per-Gaussian confidence driving pruning, densification and reduced learning rates for stable Gaussians","none beyond the motion prior (pose history or IMU); no BIM or scene prior","3D Gaussian map with rendered RGB and depth and a reconstructed point cloud; geometry assessed qualitatively and by cloud-to-mesh distance ranges on Replica (Fig. 15), no tabulated geometric accuracy","NVIDIA RTX 4090 (paper states 48 GB VRAM) + Intel Core i9-13900K; 1.09 to 1.84 FPS on Replica, 1.08 FPS for the full system on TUM fr1_desk; 4.63 GB GPU memory; tracking 15.85 ms and mapping 88 ms per iteration",null,"not_verified",[],{"id":5,"kind":52,"shortName":7,"title":8,"authors":53,"year":9,"venue":56,"venueType":57,"publisher":58,"volumeIssuePages":59,"doi":60,"arxivId":48,"url":61,"firstPublicDate":62,"publicationStatus":16,"metadataStatus":63,"fulltextStatus":15,"era":10,"classicReason":64,"codeUrl":48,"cluster":11,"topics":65,"mdpi":66,"verification":67,"label":6,"fulltextRoute":68,"versionRead":69,"addedByCensus":66},"method",[54,55],"Yuanyuan Deng","Vincent J.L. Gan","Advanced Engineering Informatics","journal","Elsevier","71, 104317","10.1016\u002Fj.aei.2026.104317","https:\u002F\u002Fapi.crossref.org\u002Fworks\u002F10.1016\u002Fj.aei.2026.104317","2026-01-14","metadata_verified","not_applicable",[11],false,"corrected","NTU institutional (Chrome)","version of record, Advanced Engineering Informatics 71 Part B (April 2026) 104317, ScienceDirect HTML full text",[71,77,81,87],{"category":72,"model":73,"canonical":73,"role":74,"dataset":48,"specs":75,"locator":76},"compute","NVIDIA RTX 4090","compute for runtime","stated as 48 GB VRAM (Sec. 4.1); MonoGS reproduced on it (Table 8 note); Orbeez-SLAM, Point-SLAM and SplaTAM FPS cited from GS-ICP-SLAM, which also used an RTX 4090; hardware for the Table 9 baseline memory values not stated","Sec. 4.1; Table 8 note",{"category":72,"model":78,"canonical":78,"role":74,"dataset":48,"specs":79,"locator":80},"Intel Core i9-13900K","desktop CPU","Sec. 4.1",{"category":82,"model":83,"canonical":83,"role":84,"dataset":85,"specs":86,"locator":80},"rgbd","Microsoft Kinect","dataset sensor","TUM RGB-D","sensor of the TUM real-world indoor sequences",{"category":88,"model":89,"canonical":89,"role":90,"dataset":85,"specs":91,"locator":80},"other","motion capture system (model not reported)","reference or ground truth","ground-truth trajectories of TUM",[],{"totalRows":94,"groupCount":95,"groups":96,"others":539},24,6,[97,173,235,381],{"slug":98,"group":99,"sourceId":5,"sourceLabel":6,"table":100,"selfRows":101,"metrics":102,"seqs":108,"entrants":125,"cells":132,"outcomes":167,"locators":168,"hardware":169,"wordings":170,"notes":171},"deng2026-mcgs-slam-table-3","deng2026_mcgs_slam:Table 3","Table 3",7,[103],{"label":104,"unit":105,"statistic":106,"alignment":107},"ATE RMSE (cm)","cm","RMSE","not_reported",[109,113,115,117,119,121,123],{"dataset":110,"sequence":111,"environment":112},"Replica","office0","synthetic indoor",{"dataset":110,"sequence":114,"environment":112},"office1",{"dataset":110,"sequence":116,"environment":112},"office2",{"dataset":110,"sequence":118,"environment":112},"office3",{"dataset":110,"sequence":120,"environment":112},"room0",{"dataset":110,"sequence":122,"environment":112},"room1",{"dataset":110,"sequence":124,"environment":112},"room2",[126,130],{"name":127,"methodId":128,"linkable":129,"proposed":66,"self":66},"Baseline MonoGS","monogs2024",true,{"name":131,"methodId":5,"linkable":129,"proposed":129,"self":129},"MCGS-SLAM (Ours)",[133,137,140,142,144,147,149,152,153,156,158,161,163,165],[134,134,134,135,136,134,136,136,134],0,0.444,-1,[138,134,134,139,136,134,136,136,134],1,0.42,[134,134,138,141,136,134,136,136,134],0.594,[138,134,138,143,136,134,136,136,134],0.557,[134,134,145,146,136,134,136,136,134],2,0.206,[138,134,145,148,136,134,136,136,134],0.211,[134,134,150,151,136,134,136,136,134],3,0.179,[138,134,150,146,136,134,136,136,134],[134,134,154,155,136,134,136,136,134],4,0.422,[138,134,154,157,136,134,136,136,134],0.437,[134,134,159,160,136,134,136,136,134],5,1.952,[138,134,159,162,136,134,136,136,134],0.33,[134,134,95,164,136,134,136,136,134],0.34,[138,134,95,166,136,134,136,136,134],0.328,[],[100],[],[],[172],"Replica pose tracking, baseline MonoGS versus MCGS-SLAM",{"slug":174,"group":175,"sourceId":5,"sourceLabel":6,"table":176,"selfRows":101,"metrics":177,"seqs":181,"entrants":196,"cells":200,"outcomes":228,"locators":229,"hardware":230,"wordings":232,"notes":233},"deng2026-mcgs-slam-table-8","deng2026_mcgs_slam:Table 8","Table 8",[178],{"label":179,"unit":180,"statistic":107,"alignment":64},"FPS","Hz",[182,184,186,188,190,192,194],{"dataset":110,"sequence":183,"environment":112},"O0",{"dataset":110,"sequence":185,"environment":112},"O1",{"dataset":110,"sequence":187,"environment":112},"O2",{"dataset":110,"sequence":189,"environment":112},"O3",{"dataset":110,"sequence":191,"environment":112},"R0",{"dataset":110,"sequence":193,"environment":112},"R1",{"dataset":110,"sequence":195,"environment":112},"R2",[197,199],{"name":198,"methodId":128,"linkable":129,"proposed":66,"self":66},"MonoGS (re-run by authors)",{"name":131,"methodId":5,"linkable":129,"proposed":129,"self":129},[201,203,204,206,208,210,212,214,216,218,220,222,224,226],[134,134,134,202,136,134,134,136,134],1.5,[138,134,134,202,136,134,134,136,134],[134,134,138,205,136,134,134,136,134],1.85,[138,134,138,207,136,134,134,136,134],1.84,[134,134,145,209,136,134,134,136,134],1.23,[138,134,145,211,136,134,134,136,134],1.22,[134,134,150,213,136,134,134,136,134],1.14,[138,134,150,215,136,134,134,136,134],1.16,[134,134,154,217,136,134,134,136,134],1.04,[138,134,154,219,136,134,134,136,134],1.09,[134,134,159,221,136,134,134,136,134],1.34,[138,134,159,223,136,134,134,136,134],1.31,[134,134,95,225,136,134,134,136,134],1.24,[138,134,95,227,136,134,134,136,134],1.3,[],[176],[231],"NVIDIA RTX 4090 + Intel Core i9-13900K",[],[234],"Replica FPS; MonoGS reproduced on the same RTX 4090 (Orbeez-SLAM, Point-SLAM and SplaTAM columns cited from GS-ICP-SLAM, not transcribed)",{"slug":236,"group":237,"sourceId":5,"sourceLabel":6,"table":238,"selfRows":159,"metrics":239,"seqs":243,"entrants":256,"cells":285,"outcomes":374,"locators":376,"hardware":377,"wordings":378,"notes":379},"deng2026-mcgs-slam-table-11","deng2026_mcgs_slam:Table 11","Table 11",[240],{"label":241,"unit":242,"statistic":106,"alignment":107},"ATE RMSE [m]","m",[244,248,250,252,254],{"dataset":245,"sequence":246,"environment":247},"EuRoC MAV","MH01 Easy","large-scale indoor sequences with fast camera motion (Sec. 4.6)",{"dataset":245,"sequence":249,"environment":247},"MH02 Easy",{"dataset":245,"sequence":251,"environment":247},"MH03 Medium",{"dataset":245,"sequence":253,"environment":247},"MH04 Difficult",{"dataset":245,"sequence":255,"environment":247},"MH05 Difficult",[257,260,263,266,268,271,273,276,279,281,283],{"name":258,"methodId":259,"linkable":129,"proposed":66,"self":66},"DSO (monocular, w\u002Fo loop)","dso2018",{"name":261,"methodId":262,"linkable":129,"proposed":66,"self":66},"SVO (monocular, w\u002Fo loop)","svo2017",{"name":264,"methodId":265,"linkable":129,"proposed":66,"self":66},"ORB-SLAM (monocular, with loop)","orbslam2015",{"name":267,"methodId":48,"linkable":66,"proposed":66,"self":66},"DSM (monocular, w\u002Fo loop)",{"name":269,"methodId":270,"linkable":129,"proposed":66,"self":66},"VINS-Fusion (stereo, with loop)","vinsfusion2019",{"name":272,"methodId":262,"linkable":129,"proposed":66,"self":66},"SVO (stereo, w\u002Fo loop)",{"name":274,"methodId":275,"linkable":129,"proposed":66,"self":66},"ORB-SLAM3 (stereo, with loop)","orbslam3_2021",{"name":277,"methodId":278,"linkable":129,"proposed":66,"self":66},"SplaTAM (w\u002Fo loop)","splatam2024",{"name":280,"methodId":48,"linkable":66,"proposed":66,"self":66},"GI-SLAM (w\u002Fo loop)",{"name":282,"methodId":128,"linkable":129,"proposed":66,"self":66},"Baseline MonoGS (w\u002Fo loop)",{"name":284,"methodId":5,"linkable":129,"proposed":129,"self":129},"MCGS-SLAM (Ours, w\u002Fo loop)",[286,288,289,291,293,295,297,299,301,303,305,307,308,309,311,313,315,316,317,318,319,321,323,324,326,328,329,330,332,333,334,336,338,339,341,342,344,345,347,348,349,351,352,353,354,355,357,359,361,363,365,367,368,370,372],[134,134,134,287,136,134,136,136,134],0.05,[134,134,138,287,136,134,136,136,134],[134,134,145,290,136,134,136,136,134],0.17,[134,134,150,292,136,134,136,136,134],3.81,[134,134,154,294,136,134,136,136,134],0.11,[138,134,134,296,136,134,136,136,134],0.1,[138,134,138,298,136,134,136,136,134],0.12,[138,134,145,300,136,134,136,136,134],0.41,[138,134,150,302,136,134,136,136,134],0.43,[138,134,154,304,136,134,136,136,134],0.3,[145,134,134,306,136,134,136,136,134],0.07,[145,134,138,306,136,134,136,136,134],[145,134,145,306,136,134,136,136,134],[145,134,150,310,136,134,136,136,134],0.08,[145,134,154,312,136,134,136,136,134],0.06,[150,134,134,314,136,134,136,136,134],0.04,[150,134,138,314,136,134,136,136,134],[150,134,145,312,136,134,136,136,134],[150,134,150,312,136,134,136,136,134],[150,134,154,306,136,134,136,136,134],[154,134,134,320,136,134,136,136,134],0.54,[154,134,138,322,136,134,136,136,134],0.46,[154,134,145,162,136,134,136,136,134],[154,134,150,325,136,134,136,136,134],0.78,[154,134,154,327,136,134,136,136,134],0.5,[159,134,134,314,136,134,136,136,134],[159,134,138,306,136,134,136,136,134],[159,134,145,331,136,134,136,136,134],0.27,[159,134,150,290,136,134,136,136,134],[159,134,154,298,136,134,136,136,134],[95,134,134,335,136,134,136,136,134],0.03,[95,134,138,337,136,134,136,136,134],0.02,[95,134,145,337,136,134,136,136,134],[95,134,150,340,136,134,136,136,134],0.09,[95,134,154,287,136,134,136,136,134],[101,134,134,343,136,134,136,136,134],0.39,[101,134,138,48,134,134,136,136,134],[101,134,145,346,136,134,136,136,134],4.78,[101,134,150,48,134,134,136,136,134],[101,134,154,48,134,134,136,136,134],[350,134,134,296,136,134,136,136,134],8,[350,134,138,306,136,134,136,136,134],[350,134,145,48,134,134,136,136,134],[350,134,150,48,134,134,136,136,134],[350,134,154,48,134,134,136,136,134],[356,134,134,298,136,134,136,136,134],9,[356,134,138,358,136,134,136,136,134],0.14,[356,134,145,360,136,134,136,136,134],2.18,[356,134,150,362,136,134,136,136,134],4.52,[356,134,154,364,136,134,136,136,134],3.19,[366,134,134,287,136,134,136,136,134],10,[366,134,138,306,136,134,136,136,134],[366,134,145,369,136,134,136,136,134],1.79,[366,134,150,371,136,134,136,136,134],2.25,[366,134,154,373,136,134,136,136,134],3.78,[375],"not_reported (dash in table)",[238],[],[],[380],"EuRoC MH01 to MH05; input modality for the 3DGS methods not stated; classical-method values match those listed in DROID-SLAM Tables 3 and 5 after rounding",{"slug":382,"group":383,"sourceId":5,"sourceLabel":6,"table":384,"selfRows":150,"metrics":385,"seqs":387,"entrants":395,"cells":434,"outcomes":533,"locators":534,"hardware":535,"wordings":536,"notes":537},"deng2026-mcgs-slam-table-1","deng2026_mcgs_slam:Table 1","Table 1",[386],{"label":104,"unit":105,"statistic":106,"alignment":107},[388,391,393],{"dataset":85,"sequence":389,"environment":390},"freiburg1_desk","real-world indoor sequences captured by Microsoft Kinect sensors (TUM RGB-D)",{"dataset":85,"sequence":392,"environment":390},"freiburg2_xyz",{"dataset":85,"sequence":394,"environment":390},"freiburg3_long_office_household",[396,398,400,402,404,406,408,410,412,415,417,419,421,424,427,429,432,433],{"name":397,"methodId":278,"linkable":129,"proposed":66,"self":66},"SplaTAM",{"name":399,"methodId":48,"linkable":66,"proposed":66,"self":66},"MM3DGS-SLAM",{"name":401,"methodId":275,"linkable":129,"proposed":66,"self":66},"ORB-SLAM3",{"name":403,"methodId":48,"linkable":66,"proposed":66,"self":66},"CVO-SLAM",{"name":405,"methodId":48,"linkable":66,"proposed":66,"self":66},"RKD-SLAM",{"name":407,"methodId":48,"linkable":66,"proposed":66,"self":66},"CG-SLAM",{"name":409,"methodId":48,"linkable":66,"proposed":66,"self":66},"GS-ICP-SLAM",{"name":411,"methodId":48,"linkable":66,"proposed":66,"self":66},"GLC-SLAM",{"name":413,"methodId":414,"linkable":129,"proposed":66,"self":66},"NICE-SLAM","niceslam2022",{"name":416,"methodId":48,"linkable":66,"proposed":66,"self":66},"Uni-SLAM",{"name":418,"methodId":48,"linkable":66,"proposed":66,"self":66},"DI-Fusion (reference [57] is DI-SLAM)",{"name":420,"methodId":48,"linkable":66,"proposed":66,"self":66},"Vox-Fusion",{"name":422,"methodId":423,"linkable":129,"proposed":66,"self":66},"ESLAM","eslam2023",{"name":425,"methodId":426,"linkable":129,"proposed":66,"self":66},"Co-SLAM","coslam2023",{"name":428,"methodId":48,"linkable":66,"proposed":66,"self":66},"Gaussian-SLAM",{"name":430,"methodId":431,"linkable":129,"proposed":66,"self":66},"Point-SLAM","pointslam2023",{"name":127,"methodId":128,"linkable":129,"proposed":66,"self":66},{"name":131,"methodId":5,"linkable":129,"proposed":129,"self":129},[435,437,438,440,442,444,445,447,448,450,452,453,454,456,458,460,462,463,465,467,469,470,471,472,474,476,478,480,482,484,486,488,489,491,494,496,498,501,503,505,507,508,509,512,514,516,518,519,521,524,526,528,531,532],[134,134,134,436,136,134,136,136,134],3.35,[134,134,138,225,136,134,136,136,134],[134,134,145,439,136,134,136,136,134],5.16,[138,134,134,441,136,134,136,136,134],3.51,[138,134,138,443,136,134,136,136,134],2.04,[138,134,145,48,134,134,136,136,134],[145,134,134,446,136,134,136,136,134],2.07,[145,134,138,343,136,134,136,136,134],[145,134,145,449,136,134,136,136,134],1.7,[150,134,134,451,136,134,136,136,134],3.15,[150,134,138,48,134,134,136,136,134],[150,134,145,48,134,134,136,136,134],[154,134,134,455,136,134,136,136,134],2.1,[154,134,138,457,136,134,136,136,134],1.2,[154,134,145,459,136,134,136,136,134],2.8,[159,134,134,461,136,134,136,136,134],2.4,[159,134,138,457,136,134,136,136,134],[159,134,145,464,136,134,136,136,134],2.5,[95,134,134,466,136,134,136,136,134],2.7,[95,134,138,468,136,134,136,136,134],1.8,[95,134,145,466,136,134,136,136,134],[101,134,134,205,136,134,136,136,134],[101,134,138,227,136,134,136,136,134],[101,134,145,473,136,134,136,136,134],3.53,[350,134,134,475,136,134,136,136,134],4.3,[350,134,138,477,136,134,136,136,134],3.17,[350,134,145,479,136,134,136,136,134],3.9,[356,134,134,481,136,134,136,136,134],2.37,[356,134,138,483,136,134,136,136,134],1.17,[356,134,145,485,136,134,136,136,134],2.62,[366,134,134,487,136,134,136,136,134],4.4,[366,134,138,145,136,134,136,136,134],[366,134,145,490,136,134,136,136,134],5.8,[492,134,134,493,136,134,136,136,134],11,3.52,[492,134,138,495,136,134,136,136,134],1.49,[492,134,145,497,136,134,136,136,134],26.01,[499,134,134,500,136,134,136,136,134],12,2.47,[499,134,138,502,136,134,136,136,134],1.11,[499,134,145,504,136,134,136,136,134],2.42,[506,134,134,461,136,134,136,136,134],13,[506,134,138,449,136,134,136,136,134],[506,134,145,461,136,134,136,136,134],[510,134,134,511,136,134,136,136,134],14,2.73,[510,134,138,513,136,134,136,136,134],1.39,[510,134,145,515,136,134,136,136,134],5.31,[517,134,134,475,136,134,136,136,134],15,[517,134,138,223,136,134,136,136,134],[517,134,145,520,136,134,136,136,134],3.48,[522,134,134,523,136,134,136,136,134],16,2.13,[522,134,138,525,136,134,136,136,134],1.44,[522,134,145,527,136,134,136,136,134],1.51,[529,134,134,530,136,134,136,136,134],17,1.52,[529,134,138,513,136,134,136,136,134],[529,134,145,525,136,134,136,136,134],[375],[384],[],[],[538],"TUM static scenes; ATE RMSE; the paper does not state whether baseline values were re-run or taken from the literature",[540,545],{"group":541,"slug":542,"sourceLabel":6,"table":543,"selfRows":138,"datasets":544},"deng2026_mcgs_slam:Table 2","deng2026-mcgs-slam-table-2","Table 2",[85],{"group":546,"slug":547,"sourceLabel":6,"table":548,"selfRows":138,"datasets":549},"deng2026_mcgs_slam:Table 9","deng2026-mcgs-slam-table-9","Table 9",[85],1790510665388]