Research map / 3D vision / Depth and stereo
VGGT-Motion: Motion-Aware Calibration-Free Monocular SLAM for Long-Range Consistency
Zhuang Xiong, Chen Zhang, Qingshan Xu, Wenbing Tao
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 parallax-based keyframe selection instead of dense sampling (ablation) | VGGT-Motion · KITTI (mean ATE excl. seq 01) | ATE (m): 48.67 -> 27.21 | Improved |
| use parallax-based keyframe selection instead of uniform sampling (ablation) | VGGT-Motion · KITTI (mean ATE excl. seq 01) | ATE (m): 34.14 -> 27.21 | Improved |
| use parallax-based instead of covisibility-based keyframe selection (ablation) | VGGT-Motion · KITTI (mean ATE excl. seq 01) | ATE (m): 33.81 -> 27.21 | Improved |
| add static redundancy pruning (ablation) | parallax-based keyframe selection · KITTI (mean ATE excl. seq 01) | ATE (m): 27.21 -> 26.98 | Improved |
| add static redundancy pruning (ablation) | parallax-based keyframe selection · KITTI stop-and-go sequences (ATE*) | ATE* (m): 15.92 -> 15.36 | Improved |
| replace temporal slicing with topology-aware partitioning (turning encapsulation) (ablation) | VGGT-Motion submap construction · KITTI | ATE (m): 26.98 -> 24.56 | Improved |
| replace temporal slicing with topology-aware partitioning (ablation) | VGGT-Motion submap construction · KITTI | translation drift (%): 1.58 -> 1.41 | Improved |
| use cumulative-parallax triggered partitioning instead of temporal slicing (ablation) | VGGT-Motion submap construction · KITTI | ATE (m): 26.98 -> 28.15 | Hurt |
| use anchor-driven direct Sim(3) registration instead of dense overlap alignment (ablation) | VGGT-Motion submap alignment · KITTI | ATE and alignment time | Improved |
| use VGGT-Motion instead of VGGT-Long on KITTI | VGGT-Long · KITTI | ATE (m): 27.64 -> 24.17 | Improved |
| use VGGT-Motion instead of VGGT-Long on Waymo | VGGT-Long · Waymo Open | ATE (m): 2.00 -> 1.59 | Improved |
| use VGGT-Motion instead of VGGT-Long on long zero-shot sequences | VGGT-Long · 4Seasons / Complex Urban / A2D2 | ATE (m): 280.25 -> 12.22 (4Seasons) | Improved |
Closest papers by meaning
- VGGT-SLAM: Dense RGB SLAM Optimized on the SL(4) Manifold (NeurIPS 2025 · similarity 0.91)
- Robust and Efficient Monocular 3D Gaussian SLAM for Kilometer-Scale Outdoor Scenes (ECCV 2026 · similarity 0.88)
- UniSim-SLAM: Feed-Forward SLAM with Unified Sim(3) Optimization (ECCV 2026 · similarity 0.88)
- DG-SLAM: Robust Dynamic Gaussian Splatting SLAM with Hybrid Pose Optimization (NeurIPS 2024 · similarity 0.87)
- Gaussian Splatting SLAM (CVPR 2024 · similarity 0.86)
- WildGS-SLAM: Monocular Gaussian Splatting SLAM in Dynamic Environments (CVPR 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.
