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
Thanks for sharing your excellent work! It seems that the way of resizing GT in your code is nearest neighbor. And these is no scale jittering augmentation in your implementation. When gt is dense, nearest neighbor can lead to less flying point in object boundary but less precise than bilinear interpolation in other area. I am not clear what influence bilinear interpolation can bring to your method, especially to the boundary.
The text was updated successfully, but these errors were encountered:
Thanks for sharing your excellent work! It seems that the way of resizing GT in your code is nearest neighbor. And these is no scale jittering augmentation in your implementation. When gt is dense, nearest neighbor can lead to less flying point in object boundary but less precise than bilinear interpolation in other area. I am not clear what influence bilinear interpolation can bring to your method, especially to the boundary.
The text was updated successfully, but these errors were encountered: