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
What I'm most interested is by how much the 3D convolution improved the result?
Using a single module on the volume, rather than multiple
modules on each of the confidence maps, is more efficient
and leads to better results since it can use information from
the neighboring disparity planes.
Hi,
As the CVPR is finished, when do you plan to open-source the code?
Thanks.
Bests
The text was updated successfully, but these errors were encountered: