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
However, the MORAN is not designed for vertical text. Our method was proposed for the complicated deformation of text within a cropped horizontal rectangle.
Would you be able to elaborate on why this is, and if it would be trivial to adjust the network to support vertical text.
Cheers.
The text was updated successfully, but these errors were encountered:
Thanks for your attention. One potential solution is to rotate the images (90 degree or -90 degree) of both the training and testing sets. Our new work DAN (2D attention) is also worth a try.
As stated in the end of the paper:
Would you be able to elaborate on why this is, and if it would be trivial to adjust the network to support vertical text.
Cheers.
The text was updated successfully, but these errors were encountered: