tensorflow-stuff

/testimgs: images of 'fuel' object for testing

cvc.py: quick tool to cleanly destroy all cv2 windows

imgrn.py: should create random image names to avoid file name conflicts. not sure if working

keras_model.h5: pretrained image recognition ML model

labels.txt: category labels for it

ml_test_from_image.py: should open a specified image and run it through the ML model, for repeatable/convenient testing purpose

testimg.jpg: red ball image, model never trained on

tffull.py: master file, with pi camera integration

tffull

with live view on screen, press c to capture image and send to ML