ML training datasets typically need to be a a square, and a multiple of a base in size (ex. 32x32, 64x64).
This tool is built to quickly (manually) crop images to fit that specificaiton.
- Space key = crop image and save
- s key = skip image
- q key = quit
Wrote all the code on windows, but the code should be OS independent. Let me know if it fails to run in linux.