we were wondering whether one could save one cropped image per identified object, where the cropping is the bounding box of the object maybe plus a certain margin.
so for instance, if we have one big circular object (1) and one small elongated object (2) in an image, the output should be two images
image 1: big square shape, containing object 1
image 2: small rectangular shape, containing object 2
is that possible?
Thanks a lot!