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
Please Refer to the Data Preparation section of the README file:
Note that the land-cover maps are RGB images in the original SECOND dataset for visualization, you need to transform them into single-channel. Also, the binary change maps should be generated by yourself and put them into folder [GT_CD].
当进入imutils.py脚本中的random_crop_mcd时,似乎变化前和变化后的标签维度是(512, 512, 3),而在这个方法中将其维度视作(512, 512),这样就会导致维度对应不上,请问应该如何修改呢
The text was updated successfully, but these errors were encountered: