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
Thank you for your interest! In this field, it is common practice to use a Gaussian filter. To ensure fair comparisons, we typically use the same Gaussian filter parameters.
I enjoyed ur paper and visit ur repo.
First of all, I'm new in this anomaly detection task.
https://github.com/caoyunkang/AdaCLIP/blob/main/method/trainer.py#L193-L196
What I don't understand is that anomaly map is Gaussian filtered before calculating metrics.
I think controlling this parameter may change the results.
(using Gaussian filter or not, if use then what sigma to use, ...)
Is there specific reason for this? or is it common in this field?
The text was updated successfully, but these errors were encountered: