Fix bundle adjustment performance regression due to changed Gauge #3527
+111
−77
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
sarlinpe
approved these changes
Aug 4, 2025
Member
sarlinpe
left a comment
There was a problem hiding this comment.
Wow nice find! Is the slow-down due to increased step time or increased number of iterations until convergence?
B1ueber2y
approved these changes
Aug 4, 2025
Contributor
Author
Already discussed offline, but for the record here: primarily more steps to converge. |
ahojnnes
added a commit
that referenced
this pull request
Aug 4, 2025
) Based on reported issue #3479, I root-caused a performance regression in bundle adjustment to different gauge fixing behavior. In particular, the change from fixing the gauge with two cameras vs. fixing it with three points causes a ~40% slowdown in global bundle adjustment. No significant difference can be measured in local bundle adjustment. I could also not find a significant difference in the obtained solution quality. This PR reverts back to using two cameras for fixing the gauge. If no such two cameras can be found, which can occasionally happen in case of configured non-trivial rigs, we fall back to fixing three points.
tavislocus
pushed a commit
to tavislocus/colmap_6dof
that referenced
this pull request
Aug 19, 2025
…lmap#3527) Based on reported issue colmap#3479, I root-caused a performance regression in bundle adjustment to different gauge fixing behavior. In particular, the change from fixing the gauge with two cameras vs. fixing it with three points causes a ~40% slowdown in global bundle adjustment. No significant difference can be measured in local bundle adjustment. I could also not find a significant difference in the obtained solution quality. This PR reverts back to using two cameras for fixing the gauge. If no such two cameras can be found, which can occasionally happen in case of configured non-trivial rigs, we fall back to fixing three points.
tavislocus
pushed a commit
to tavislocus/colmap_6dof
that referenced
this pull request
Aug 22, 2025
…lmap#3527) Based on reported issue colmap#3479, I root-caused a performance regression in bundle adjustment to different gauge fixing behavior. In particular, the change from fixing the gauge with two cameras vs. fixing it with three points causes a ~40% slowdown in global bundle adjustment. No significant difference can be measured in local bundle adjustment. I could also not find a significant difference in the obtained solution quality. This PR reverts back to using two cameras for fixing the gauge. If no such two cameras can be found, which can occasionally happen in case of configured non-trivial rigs, we fall back to fixing three points.
tavislocus
pushed a commit
to tavislocus/colmap_6dof
that referenced
this pull request
Aug 22, 2025
…lmap#3527) Based on reported issue colmap#3479, I root-caused a performance regression in bundle adjustment to different gauge fixing behavior. In particular, the change from fixing the gauge with two cameras vs. fixing it with three points causes a ~40% slowdown in global bundle adjustment. No significant difference can be measured in local bundle adjustment. I could also not find a significant difference in the obtained solution quality. This PR reverts back to using two cameras for fixing the gauge. If no such two cameras can be found, which can occasionally happen in case of configured non-trivial rigs, we fall back to fixing three points.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Based on reported issue #3479, I root-caused a performance regression in bundle adjustment to different gauge fixing behavior. In particular, the change from fixing the gauge with two cameras vs. fixing it with three points causes a ~40% slowdown in global bundle adjustment. No significant difference can be measured in local bundle adjustment. I could also not find a significant difference in the obtained solution quality. This PR reverts back to using two cameras for fixing the gauge. If no such two cameras can be found, which can occasionally happen in case of configured non-trivial rigs, we fall back to fixing three points.