LongDPM: Overlap-Aware 4D Reconstruction from Long Monocular Videos
For researchers in dynamic 3D reconstruction, LongDPM addresses the scalability limitation of feed-forward models to long videos while maintaining dense geometry and correspondence.
LongDPM enables consistent 4D reconstruction from long monocular videos by processing overlapping chunks, registering them with confidence-weighted alignment, and fusing dynamic trajectories. It reduces dense tracking EPE over V-DPM on PointOdyssey, Kubric-F, and Kubric-G, and achieves best TUM-dynamics ATE for camera pose estimation.
Recovering a dynamic 3D scene from a long monocular video is crucial for dense geometry, camera motion, and temporal correspondence to remain consistent in a shared coordinate system. Existing methods face two key challenges: (1) feed-forward reconstruction models provide accurate local predictions but are limited to short clips, and (2) long-range trackers preserve correspondences without producing dense sequence-level reconstruction. This paper presents LongDPM, a novel overlap-aware framework for scalable long-range monocular dynamic reconstruction. First, LongDPM processes long videos in overlapping chunks, keeping inference memory bounded by the chunk length. Second, it connects chunk-local coordinate systems through confidence-weighted registration with static-aware overlap abstraction. Third, it associates dynamic identities across chunk boundaries and fuses matched trajectories to recover coherent long-range 3D motion. Experimental results demonstrate that LongDPM achieves superior long-range reconstruction and tracking performance, reducing dense tracking EPE over V-DPM on PointOdyssey, Kubric-F, and Kubric-G, while obtaining the best TUM-dynamics ATE for camera pose estimation.