I was wondering as to what would be my best approach to this problem. I have a 6000-by-6000 image with a 6000-by-6000 mask. I wanted to crop the image into several sub-images before training and came across Extract_Patches_2d in Scikit learn. Looks like the tool to get the job done but I have one issue. If I run this on a single image, how can I be sure that it will use the same patch as a blueprint for the image mask as well ?
Asked
Active
Viewed 259 times
