[{"data":1,"prerenderedAt":102},["ShallowReactive",2],{"method-voxfusion2022":3},{"method":4,"reference":58,"equipment":82,"figures":101,"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":34,"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},"voxfusion2022","Yang et al., 2022","Vox-Fusion","Vox-Fusion: Dense Tracking and Mapping with Voxel-based Neural Implicit Representation",2022,"recent","C09","odometry_with_local_mapping","Vox-Fusion 將神經隱式表面與傳統體素融合結合：場景以八元樹（octree）與 Morton 編碼管理的稀疏體素表示，體素頂點存放共享的特徵向量，再由多層感知器解碼成 SDF 與顏色。新影格的深度點雲一旦落在既有體素之外就即時配置新體素，因此不必預先知道場景範圍，記憶體也只花在有觀測的表面附近。追蹤時固定地圖，只以可微分體積渲染最佳化相機位姿；建圖時則對隨機挑選的關鍵影格視窗聯合最佳化地圖與位姿，但系統沒有迴圈閉合或全域最佳化。","RGB-D neural implicit SLAM whose SDF map is stored as embeddings on a sparse, dynamically allocated octree of voxels, so it needs no predefined scene bound; tracking and windowed joint mapping both use differentiable SDF volume rendering, without loop closure.","full_text_reviewed","peer_reviewed_published","main_body","論文未涉及營建場域；定量評估只在 Replica 合成場景與 ScanNet 五個室內場景進行，iPhone 13 Pro 與 iPad Pro 的室外手持掃描僅作定性展示。其稀疏體素依觀測動態擴張、不需預設場景邊界的設計，較適合範圍事先未知的工地（推論），但系統無迴圈閉合，作者也承認長時間追蹤會漂移，大範圍施工現場仍需額外的全域校正。",[20,21],"public_benchmark","simulation",[23,24,25,26],"Dynamic sparse voxel allocation removes the need for a known scene bound, unlike NICE-SLAM's pre-allocated dense grid (Sec. 2.1, Sec. 4.4)","Embedding memory on Replica office-0 is 0.149 MB versus 238.88 MB for NICE-SLAM (Table 5)","Replica average ATE RMSE 0.0054 m versus 0.0195 m for NICE-SLAM, and average accuracy 2.37 cm, completion 2.28 cm, completion ratio 92.86% (Tables 1-2)","Leaves large unobserved regions empty instead of hallucinating surfaces, so observed and unobserved space remain distinguishable (Sec. 5.2)",[28,29,30,31],"Cannot robustly handle dynamic objects or drift in long-time tracking (Sec. 7)","No loop closure or global optimization (Sec. 4.3, Sec. 7) (inference from the method description)","Typical rates are about 5 Hz tracking and 2 Hz mapping on an RTX 3090, and may be slower in challenging scenes (Sec. 5.4)","Low-resolution iOS depth limits reconstruction quality of the handheld outdoor and indoor captures, which are shown only qualitatively (Sec. 5.3, Fig. 8)",[33],"RGB-D camera (synthetic Replica, ScanNet, iPhone 13 Pro and iPad Pro (2020) with onboard LiDAR depth)",[35,36,37],"simulation (Replica synthetic RGB-D sequences)","handheld (iPhone 13 Pro and iPad Pro captures, Fig. 8)","not_reported (ScanNet capture platform not described)","gradient-based 6-DoF pose optimization in se(3) through differentiable SDF volume rendering against a frozen copy of the map (zero-motion initialization); mapping jointly optimizes decoder, voxel embeddings and poses of a random keyframe window (Sec. 4.2-4.3)","direct: rendered colour and depth losses plus free-space and SDF losses on sparsely sampled pixels whose rays hit allocated voxels (Sec. 4.1)","discrete poses","not_applicable","none (the authors list drift in long-time tracking as unresolved, Sec. 7)","none; joint optimization only over a window of randomly selected keyframes (Sec. 4.3)","sparse voxel grid (voxel size 0.2 m) in an octree with Morton coding, 16-D embeddings on voxel vertices shared by neighbours, decoded by an MLP into SDF and colour; voxels allocated on the fly from back-projected depth, so no scene bound is needed (Sec. 4.4, Sec. 5.1)","none; decoder and embeddings are learned on the fly without pre-trained geometry priors (Sec. 2.1)","camera trajectory and SDF-based surface mesh (evaluated as mesh against ground-truth mesh); rendered colour and depth images (Sec. 5)","single NVIDIA RTX 3090; about 150-200 ms to track a frame and 450-550 ms per joint optimization, i.e. about 5 Hz tracking and 2 Hz mapping (Sec. 5.4, Table 4)","https:\u002F\u002Fgithub.com\u002Fzju3dv\u002FVox-Fusion","not_reported (no LICENSE file in the repository; GitHub reports no license)",[51,55],{"relation":52,"title":53,"doi_or_url":54},"preprint","arXiv:2210.15858v3","https:\u002F\u002Farxiv.org\u002Fabs\u002F2210.15858",{"relation":56,"title":57,"doi_or_url":48},"code_release","zju3dv\u002FVox-Fusion",{"id":5,"kind":59,"shortName":7,"title":8,"authors":60,"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":41,"codeUrl":48,"cluster":11,"topics":76,"mdpi":77,"verification":78,"label":6,"fulltextRoute":79,"versionRead":80,"addedByCensus":81},"method",[61,62,63,64,65,66],"Xingrui Yang","Hai Li","Hongjia Zhai","Yuhang Ming","Yuqian Liu","Guofeng Zhang","2022 IEEE International Symposium on Mixed and Augmented Reality (ISMAR)","conference","IEEE","pp. 499-507","10.1109\u002Fismar55827.2022.00066","2210.15858","https:\u002F\u002Fapi.crossref.org\u002Fworks\u002F10.1109\u002FISMAR55827.2022.00066","2022-10-28","metadata_verified",[11],false,"corrected","arXiv","arXiv v3 (2210.15858v3, 6 Mar 2023), posted after the ISMAR 2022 paper and linked to its DOI; not compared line by line with the IEEE version of record",true,[83,90,94],{"category":84,"model":85,"canonical":85,"role":86,"dataset":87,"specs":88,"locator":89},"rgbd","iPhone 13 Pro","method input",null,"RGB images with depth from the onboard lidar sensor; very low depth resolution","Sec. 5.1, Sec. 5.3, Fig. 8",{"category":84,"model":91,"canonical":91,"role":86,"dataset":87,"specs":92,"locator":93},"iPad Pro (2020)","iOS device with range sensor used for RGB-D capture","Sec. 5.1, Sec. 5.3",{"category":95,"model":96,"canonical":97,"role":98,"dataset":87,"specs":99,"locator":100},"compute","NVidia RTX 3090","NVIDIA RTX 3090","compute for runtime","single video card used for profiling","Sec. 5.4",[],1790510660625]