Try following this example:
https://www.mathworks.com/help/images/single-image-super-resolution-using-deep-learning.html while this doesn't do exactly what you're looking for, you want to set up a
randomPatchExtractionDatastore
As input to your network. Put your input images and your output images to the network in two imagedatastores, and then use randomPatchExtractionDatastore(imdsIN,imdsOUT,patchSize)
This example helped me figure out how to customize an image-to-image network like you're trying to do. Let me know when you get results!
