-
Notifications
You must be signed in to change notification settings - Fork 52
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Remove correspondences from translation averaging outliers downstream #755
Conversation
Based on evaluation on larger datasets, this change produces small improvements (1% Pose AUC post BA) for some datasets (Barn, Kings college, Skansen, Gerrard hall) and small regressions of similar magnitude for others (skydio 501). this also somewhat depends on the frontend being used. Some example numbers:
more numbers are in the spreadsheet (ta-outliers-x) |
requesting high-level thoughts on whether we should proceed with this change. Shall update the code if thats the case. |
What degree threshold are the previous pose AUC statistics at? |
Most of them are at 1 deg, I used 2.5 degrees for a couple (including skydio501) because their pose AUCs at 1 deg are 0. Ive highlighted the rows in the spreadsheet that I used. |
gtsfm/multi_view_optimizer.py
Outdated
@@ -196,3 +199,16 @@ def get_2d_tracks( | |||
) -> List[SfmTrack2d]: | |||
tracks_estimator = CppDsfTracksEstimator() | |||
return tracks_estimator.run(corr_idxs_dict, keypoints_list) | |||
|
|||
|
|||
def filter_corr_by_idx(correspondences, idxs): |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nit: missing type annotations
Btw @akshay-krishnan I believe CI is failing:
|
could you pull in the latest from master? |
No description provided.