Research map / 3D vision / 3D generation
BundleSDF: Neural 6-DoF Tracking and 3D Reconstruction of Unknown Objects
Bowen Wen, Jonathan Tremblay, Valts Blukis, Stephen Tyree, Thomas Müller, Alex Evans, Dieter Fox, Jan Kautz, Stan Birchfield
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use BundleSDF instead of BundleTrack | object tracking and reconstruction · HO3D | ADD (%): 66.01 -> 92.62 | Improved |
| Use BundleSDF instead of BundleTrack | object tracking and reconstruction · YCBInEOAT | ADD (%): 87.26 -> 86.95 | Mixed |
| Use BundleSDF instead of BundleTrack | object tracking and reconstruction · BEHAVE | ADD (%): 45.03 -> 67.52 | Improved |
| Add memory modules to frame-to-frame coarse pose estimation (ablation) | BundleSDF · HO3D | ADD (%): 56.96 -> 92.62 | Improved |
| Add Neural Object Field and its pose loss (ablation) | BundleSDF · HO3D | ADD (%): 76.69 -> 92.62 | Improved |
| Use Neural Object Field instead of global pose graph optimization over all memory frames (ablation) | BundleSDF · HO3D | ADD (%): 78.82 -> 92.62 | Improved |
| Use hybrid SDF modeling rather than foreground rays only (ablation) | BundleSDF · HO3D | CD (cm): 2.62 -> 0.61 | Improved |
| Use a compact memory pool rather than the baseline frame-selection strategy (ablation) | BundleSDF · HO3D | ADD (%): 59.99 -> 92.62 | Improved |
Later papers that compared with this method
2 later accepted papers report BundleSDF as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- 6DOPE-GS: Online 6D Object Pose Estimation using Gaussian Splatting (ICCV 2025 · better on 7 of 8 shared results)
- ICON: Incremental CONfidence for Joint Pose and Radiance Field Optimization (CVPR 2024 · better on 0 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DROID-SLAM — DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Cameras (NeurIPS 2021) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Tracking by 3D Model Estimation of Unknown Objects in Videos (ICCV 2023 · similarity 0.88)
- Deep Active Contours for Real-time 6-DoF Object Tracking (ICCV 2023 · similarity 0.88)
- Free-Moving Object Reconstruction and Pose Estimation with Virtual Camera (AAAI 2025 · similarity 0.86)
- SingRef6D: Monocular Novel Object Pose Estimation with a Single RGB Reference (NeurIPS 2025 · similarity 0.86)
- ProxyPose: 6-DoF Pose Tracking via Video-to-Video Translation (NeurIPS 2026 · similarity 0.86)
- 4D3R: Motion-Aware Neural Reconstruction and Rendering of Dynamic Scenes from Monocular Videos (NeurIPS 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
