SpatialTracker: Tracking Any 2D Pixels in 3D Space
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917631384092672 |
|---|---|
| author | Xiao, Yuxi Wang, Qianqian Zhang, Shangzhan Xue, Nan Peng, Sida Shen, Yujun Zhou, Xiaowei |
| author_facet | Xiao, Yuxi Wang, Qianqian Zhang, Shangzhan Xue, Nan Peng, Sida Shen, Yujun Zhou, Xiaowei |
| contents | Recovering dense and long-range pixel motion in videos is a challenging problem. Part of the difficulty arises from the 3D-to-2D projection process, leading to occlusions and discontinuities in the 2D motion domain. While 2D motion can be intricate, we posit that the underlying 3D motion can often be simple and low-dimensional. In this work, we propose to estimate point trajectories in 3D space to mitigate the issues caused by image projection. Our method, named SpatialTracker, lifts 2D pixels to 3D using monocular depth estimators, represents the 3D content of each frame efficiently using a triplane representation, and performs iterative updates using a transformer to estimate 3D trajectories. Tracking in 3D allows us to leverage as-rigid-as-possible (ARAP) constraints while simultaneously learning a rigidity embedding that clusters pixels into different rigid parts. Extensive evaluation shows that our approach achieves state-of-the-art tracking performance both qualitatively and quantitatively, particularly in challenging scenarios such as out-of-plane rotation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_04319 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | SpatialTracker: Tracking Any 2D Pixels in 3D Space Xiao, Yuxi Wang, Qianqian Zhang, Shangzhan Xue, Nan Peng, Sida Shen, Yujun Zhou, Xiaowei Computer Vision and Pattern Recognition Recovering dense and long-range pixel motion in videos is a challenging problem. Part of the difficulty arises from the 3D-to-2D projection process, leading to occlusions and discontinuities in the 2D motion domain. While 2D motion can be intricate, we posit that the underlying 3D motion can often be simple and low-dimensional. In this work, we propose to estimate point trajectories in 3D space to mitigate the issues caused by image projection. Our method, named SpatialTracker, lifts 2D pixels to 3D using monocular depth estimators, represents the 3D content of each frame efficiently using a triplane representation, and performs iterative updates using a transformer to estimate 3D trajectories. Tracking in 3D allows us to leverage as-rigid-as-possible (ARAP) constraints while simultaneously learning a rigidity embedding that clusters pixels into different rigid parts. Extensive evaluation shows that our approach achieves state-of-the-art tracking performance both qualitatively and quantitatively, particularly in challenging scenarios such as out-of-plane rotation. |
| title | SpatialTracker: Tracking Any 2D Pixels in 3D Space |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2404.04319 |