ByteDance-Seed / ByteDance-Seed/Depth-Anything-3

Export the Colmap format result

Open
#194 0 comments 1 reaction 0 assignees View on GitHub
Dominant language
Python
Stars
6.3k
Forks
702
PR merge metrics
No merged PRs in 30d

Description

Hello! I noticed that Depth-Anything-3 can currently export data files in Colmap format. I wanted to further optimize this result using the BA optimization algorithm in the traditional Colmap algorithm, but found that it couldn't be accomplished. It is found that in the Colmap format result exported by DA3, 3D points are only seen by one observation (one feature on an image), so it does not meet the conditions for BA optimization. Does this conclusion hold true? If not, could you provide relevant explanations or guidance?

Contributor guide

No contributing guide indexed for this repository

Research direction

The issue names no files, tests, or entry points. Reproduce the exported Colmap result and inspect its 3D-point observation relationships against the requirements for bundle adjustment. Done means determining whether the export is valid for BA and documenting the explanation or guidance needed, or identifying the export change required.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
computer-vision
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
30/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.