facebookresearch / facebookresearch/co-tracker
Barely moving queries on the boundary of the frame
- Dominant language
- Jupyter Notebook
- Stars
- 5.1k
- Forks
- 389
- PR merge metrics
- No merged PRs in 30d
Description
Hi, I'm using the online mode of CoTrackerV3 for tracking points in 2D on static scenes with a moving camera. For some videos, I experience the problem of almost static point tracks close to the image boundary:
https://github.com/user-attachments/assets/b645e827-5aa5-40da-84ac-ab942d6f15e6
For another video with a few similarly placed points, all tracks evolve meaningfully:
https://github.com/user-attachments/assets/99c22a6e-119b-4bbd-bb0d-ef35247fe2aa
These are the visuals for the first sixty frames with point tracks, interleaved with step=6 for visualization purposes:
Here are the points whose position at frames 0 and 20 has not changed by more than 10 pixels:
`kpts=pred_tracks[:,:,np.linalg.norm(pred_tracks.cpu()[0, 0] - pred_tracks.cpu()[0, 20], axis=1)<10]`
At the same time, when I use grid-based point queries, the tracks look plausible:
https://github.com/user-attachments/assets/87d27082-a466-4a77-a10e-e8ae49367c0f
I wonder what can cause such behavior for keypoints obtained from a detector? I tried changing the resolution of the images, but it had no effect.
Contributor guide
Research direction
Start by reproducing the CoTrackerV3 online-mode case with detector-derived keypoints and compare it with the grid-based query behavior shown in the report. Check the first 60 frames and the frame-0/frame-20 displacement calculation; done means identifying why boundary tracks remain nearly static and documenting or fixing the cause.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- computer-vision, machine-learning
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100