Repository navigation
docs: point evaluation.md at the SfM evaluation plan - #40
Conversation
MUGSQA renders with exact poses, which makes its release usable as a benchmark for COLMAP settings rather than only as prior art on the input axes. Nothing in this project has varied a matcher or feature setting, because every comparison was paired on identical SfM output and so cancelled SfM quality out of the result. Issue #39 carries the plan and a table of what has been measured against what has not, so the next person asking does not re-derive it from the git log. Refs: #39
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 691fd0cf8d
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| experiment here has varied. Issue #39 holds that plan and the coverage table of | ||
| what has and has not been measured. |
There was a problem hiding this comment.
Use an explicit Markdown link to issue 39 here. In local previews and other renderers that do not apply GitHub's repository-aware issue autolinking, Issue #39 remains plain text, so readers cannot reach the SfM plan or coverage table that this commit is intended to reference.
Useful? React with 👍 / 👎.
1 participant
Three-sentence addition to the MUGSQA paragraph, noting that its exact poses make the release usable as a benchmark for our own COLMAP settings, and pointing at #39 for the plan and the coverage table.
Refs: #39