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
Hi, I was wondering if there was any plan to include bounding boxes or some kind of georeferencing for vision datasets. In many cases the information is available in the original dataset but is not exposed through the torchgeo interface. For example: CV4A Kenya Crop Type.
In general this information can be useful for visualization or generating different train-test splits based on location.
The text was updated successfully, but these errors were encountered:
Hi, I was wondering if there was any plan to include bounding boxes or some kind of georeferencing for vision datasets. In many cases the information is available in the original dataset but is not exposed through the torchgeo interface. For example: CV4A Kenya Crop Type.
In general this information can be useful for visualization or generating different train-test splits based on location.
The text was updated successfully, but these errors were encountered: