google-research / google-research/kubric
Point tracking: few occluded points are not labeled as occluded
- Dominant language
- Jupyter Notebook
- Stars
- 2.8k
- Forks
- 278
- PR merge metrics
- No merged PRs in 30d
Description
Hi,
First, thanks for making the great tool for point tracking.
I see that almost all occluded points are correctly labeled as occluded, but visually I still obverse few point among 1200=50 (videos) x 24 (points per video) are occluded. I'm sure these points are occluded because the foreground object has moved a large distance and the ground truth still stays the original position, like a position of the ground. I looked into the code, but still have not idea what caused this problem. Could you please give some suggestions or even solution? Thanks.
Best,
Zhihao
Contributor guide
Research direction
The issue does not name a file, test, or entry point. Start by reproducing the point-tracking case and tracing how occlusion labels are generated for points whose ground-truth positions remain behind a moved foreground object; done when those visibly occluded points are consistently labeled as occluded.
Written by the indexing model from the issue text.
Assessment
- Domain
- computer-vision, data
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100