requirement:
python 3.5
scikit-image
tensorflow
numpy
usage:
-
create 2 folder "yes" and "no" (or anything)
-
copy all intended image to "yes", and all other image to "no"
-
run:
import convert
convert.train_all_image_in_folder('yes',[0,1])
convert.train_all_image_in_folder('no',[1,0])
-
you can predict image using convert.predict_image(image) e.g. convert.predict_image(example.png), the result will be 1 if it's the intended image or 0 if isn't.