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
Applying EPro-PnP on two-stage networks could be tricky, unless you could make it end-to-end differentiable (e.g., differentiable pixel voting in PVNet).
I wonder if EPro-PnP can only be applied to one-stage networks, and if it can be applied to some two-stage pose estimation networks, such as PVNet.
The text was updated successfully, but these errors were encountered: