[{"data":1,"prerenderedAt":95},["ShallowReactive",2],{"method-goslam2023":3},{"method":4,"reference":60,"equipment":82,"figures":94,"results":87},{"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":36,"estimator":40,"association":41,"timeModel":42,"deskew":43,"loopClosure":44,"globalOptimization":45,"mapRepresentation":46,"prior":47,"outputGeometry":48,"compute":49,"codeUrl":50,"codeLicense":51,"relatedVersions":52},"goslam2023","Zhang et al., 2023b","GO-SLAM","GO-SLAM: Global Optimization for Consistent 3D Instant Reconstruction",2023,"recent","C09","full_slam_with_global_correction","GO-SLAM 以 DROID-SLAM 的學習式稠密光流與可微分稠密光束法平差（dense bundle adjustment）作為追蹤核心，在前端依光流估算的共視度偵測迴圈，並在獨立執行緒中對所有關鍵影格線上執行完整光束法平差，以抑制長序列的累積漂移。建圖端使用多解析度雜湊編碼的神經隱式 SDF，每次優先選取位姿變化最大的關鍵影格重新訓練，使重建隨全域最佳化後的位姿與深度同步更新。同一架構可接受單眼、立體或 RGB-D 影像，最後以 marching cubes 從 SDF 擷取網格。","Extends DROID-SLAM with flow-based loop closing and online full bundle adjustment, and continuously re-fits a hash-encoded neural SDF to the globally optimized keyframe poses and depths, giving globally consistent meshes from monocular, stereo or RGB-D video.","full_text_reviewed","peer_reviewed_published","main_body","論文未涉及營建場域；驗證資料為 TUM RGB-D、EuRoC、ETH3D-SLAM、ScanNet 與 Replica 等室內或飛行器資料集，幾何精度只在 Replica 合成場景以公分級 Accuracy 與 Completion 評估，未使用全測站或 TLS 等獨立參考量測。其長序列迴圈閉合與全域平差可說明神經隱式地圖要維持全域一致須仰賴傳統 SLAM 式的後端，但單眼 ScanNet 部分場景仍有數十公分的 ATE（Table 3），距營建量測所需精度仍有差距（推論）。",[20,21],"public_benchmark","simulation",[23,24,25,26],"Online loop closing and full BA reduce average ATE on 8 ScanNet RGB-D scenes from 11.59 cm (no LC, no full BA) to 7.02 cm, with speed falling from 30 to 10 FPS (Table 8)","On long monocular ScanNet sequences the average ATE is 17.59 cm against 52.60 cm for DROID-SLAM with final global BA (Table 3)","One architecture handles monocular, stereo and RGB-D input; stereo EuRoC average ATE 0.024 m equals DROID-SLAM (Table 2)","Mapping quality degrades little when skipping up to 7 of 8 frames (Replica F-score 85.56 to 84.41; ScanNet ATE 7.02 to 7.28 cm) (Table 7)",[28,29,30,31],"No limitations section; runtime is about 8 FPS on Replica with an RTX 3090 and 15.63 GB (Table 9) or up to 18 GB (Sec. 4.4) of GPU memory, so real-time use relies on frame skipping (Table 7) (inference)","Monocular accuracy remains scene dependent: ScanNet scene0465 still has 79.51 cm ATE (Table 3)","Reconstruction metrics are reported only on synthetic Replica; real-scene geometry is shown qualitatively (Sec. 4.4, Figs. 5, B-D) (inference)","Tracking depends on pretrained DROID-SLAM weights; generalisation beyond indoor and MAV datasets is not tested (Sec. 4.1) (inference)",[33,34,35],"monocular camera","stereo camera","RGB-D camera",[37,38,39],"UAV (EuRoC MAV dataset)","simulation (Replica synthetic scenes)","not_reported (capture platforms of TUM RGB-D, ETH3D-SLAM and ScanNet are not described in the paper)","DROID-SLAM tracking extended with online global optimization: RAFT-based recurrent update operator predicts dense flow and confidence, and a differentiable dense bundle adjustment (DBA) layer solves poses and per-pixel inverse depths by damped Gauss-Newton; front end optimizes a local keyframe window with loop edges, back end runs online full BA over all keyframes in a separate thread (Sec. 3.1)","dense learned optical flow with per-pixel confidence between keyframe pairs; keyframe-graph edges chosen by co-visibility measured as mean rigid flow (threshold tau_co = 25) with neighbourhood suppression (Sec. 3.1, Sec. 4.1)","discrete poses (keyframes)","not_applicable","flow-based: edges sampled from the unexplored part of the co-visibility matrix between local-window and historical keyframes; a loop is accepted after three consecutive candidates with mean flow below tau_co, then optimized by DBA (Sec. 3.1)","online full bundle adjustment over all keyframes in a back-end thread, running concurrently with tracking and loop closing (Sec. 3.1)","neural implicit SDF and colour field with multi-resolution hash encoding (16 levels, Instant-NGP style) and shallow MLPs, rendered by NeuS-style unbiased volume rendering; mapping re-trains on selected keyframes (latest two, top-10 by pose change, 10 stratified) (Sec. 3.2, Supp. A)","pretrained DROID-SLAM weights for tracking; rendering networks trained from scratch per scene (Sec. 4.1)","keyframe trajectory, per-keyframe depth and a triangle mesh extracted by marching cubes from the SDF (Sec. 4.1)","Intel Core i9-10920X 3.5 GHz CPU and NVIDIA RTX 3090 GPU; about 8 FPS on Replica RGB-D with 15.63 GB GPU memory in Table 9, while Sec. 4.4 states a maximum of 18 GB (Sec. 4.1, Sec. 4.4, Table 9)","https:\u002F\u002Fgithub.com\u002Fyoumi-zym\u002FGO-SLAM","Apache-2.0 (LICENSE file checked)",[53,57],{"relation":54,"title":55,"doi_or_url":56},"preprint","arXiv:2309.02436v1","https:\u002F\u002Farxiv.org\u002Fabs\u002F2309.02436",{"relation":58,"title":59,"doi_or_url":50},"code_release","youmi-zym\u002FGO-SLAM",{"id":5,"kind":61,"shortName":7,"title":8,"authors":62,"year":9,"venue":67,"venueType":68,"publisher":69,"volumeIssuePages":70,"doi":71,"arxivId":72,"url":73,"firstPublicDate":74,"publicationStatus":16,"metadataStatus":75,"fulltextStatus":15,"era":10,"classicReason":43,"codeUrl":50,"cluster":11,"topics":76,"mdpi":77,"verification":78,"label":6,"fulltextRoute":79,"versionRead":80,"addedByCensus":81},"method",[63,64,65,66],"Youmin Zhang","Fabio Tosi","Stefano Mattoccia","Matteo Poggi","2023 IEEE\u002FCVF International Conference on Computer Vision (ICCV)","conference","IEEE","pp. 3704-3714","10.1109\u002Ficcv51070.2023.00345","2309.02436","https:\u002F\u002Fapi.crossref.org\u002Fworks\u002F10.1109\u002FICCV51070.2023.00345","2023-09-05","metadata_verified",[11],false,"corrected","arXiv","arXiv v1 (2309.02436v1, 5 Sep 2023; only version, labelled ICCV 2023 by the authors) including the 4-page supplementary appended to the arXiv PDF; cross-checked against the ICCV 2023 CVF open-access paper and supplementary (all extracted values match)",true,[83,90],{"category":84,"model":85,"canonical":85,"role":86,"dataset":87,"specs":88,"locator":89},"compute","Intel Core i9-10920X","compute for runtime",null,"3.5 GHz CPU","Sec. 4.1",{"category":84,"model":91,"canonical":91,"role":86,"dataset":87,"specs":92,"locator":93},"NVIDIA RTX 3090","GPU; 15.63 GB peak memory used on Replica RGB-D (Table 9)","Sec. 4.1; Table 9",[],1790510662311]