You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The current blurriness detection in sfm-worker uses Laplacian Variance to detect blurriness, which works well for frames that have real backgrounds. However, any synthetic or well collected dataset will fail these bluriness checks due to have near solid backgrounds.
If we could find a way to dynamically switch between blur detection methods or implement a different one, that would be quite beneficial.
The text was updated successfully, but these errors were encountered:
The current blurriness detection in sfm-worker uses Laplacian Variance to detect blurriness, which works well for frames that have real backgrounds. However, any synthetic or well collected dataset will fail these bluriness checks due to have near solid backgrounds.
If we could find a way to dynamically switch between blur detection methods or implement a different one, that would be quite beneficial.
The text was updated successfully, but these errors were encountered: