We propose a novel algorithm for accelerating dense long-term 3D point tracking in videos. Through analysis of existing state-of-the-art methods, we identify two major computational bottlenecks. First, transformer-based iterative tracking becomes expensive when handling a large number of trajectories. To address this, we introduce a coarse-to-fine strategy that begins tracking with a small subset of points and progressively expands the set of tracked trajectories. The newly added trajectories are initialized using a learnable interpolation module, which is trained end-to-end alongside the tracking network. Second, we propose an optimization that significantly reduces the cost of correlation feature computation, another key bottleneck in prior methods. Together, these improvements lead to a 5-100x speedup over existing approaches while maintaining state-of-the-art tracking accuracy.
@article{arxiv.2508.01170,
title = {DELTAv2: Accelerating Dense 3D Tracking},
author = {Tuan Duc Ngo and Ashkan Mirzaei and Guocheng Qian and Hanwen Liang and Chuang Gan and Evangelos Kalogerakis and Peter Wonka and Chaoyang Wang},
journal= {arXiv preprint arXiv:2508.01170},
year = {2025}
}