Overview

3D point tracking follows points on objects or across whole scenes through video, recovering their motion in three dimensions rather than only in image coordinates. Methods combine visual correspondence with depth, multi-view geometry, point-cloud motion, and temporal constraints; newer approaches also adapt video foundation models, diffusion transformers, and state-space models for feed-forward or long-range tracking. The task supports dynamic-scene reconstruction, motion analysis, and robot learning, and is evaluated on benchmarks such as TAPVid-3D, PointOdyssey, KITTI, and nuScenes. A central challenge is maintaining accurate geometry and point identity through occlusion and long sequences without making inference too slow or memory-intensive.

Latest papers

CardsList

36 papers5 papers in the last four weeks, down 44% on the four weeks before. 0.0% of all new papers.

  1. DRoPS: Dynamic 3D Reconstruction of Pre-Scanned Objects

    Mar 25, 2026Narek Tumanyan, Samuel Rota Bulò, Denis Rozumny +5Dynamic Scene ReconstructionDeformable Shape Reconstruction

  2. 4DGS360: 360° Gaussian Reconstruction of Dynamic Objects from a Single Video

    Mar 23, 2026Jae Won Jang, Yeonjin Chang, Wonsik Shin +2Dynamic Scene ReconstructionPanoramic 3D Reconstruction

  3. PCSTracker: Long-Term Scene Flow Estimation for Point Cloud Sequences

    Mar 20, 2026Min Lin, Gangwei Xu, Xianqi Wang +2Scene Flow Estimation3D Point Tracking

  4. TrackDeform3D: Markerless and Autonomous 3D Keypoint Tracking and Dataset Collection for Deformable Objects

    Mar 17, 2026Yeheng Zong, Yizhou Chen, Alexander Bowler +23D Point Tracking

  5. 3PoinTr: From Human Videos to Robot Policies with 3D Point-Track Plans

    Mar 9, 2026Adam Hung, Bardienus Pieter Duisterhof, Jeffrey IchnowskiHuman-to-Robot Motion ImitationRobot Policy Learning

  6. Track4World: Feedforward World-centric Dense 3D Tracking of All Pixels

    Mar 3, 2026Jiahao Lu, Jiayi Xu, Wenbo Hu +5Scene Flow Estimation3D Point Tracking