My application is accident-avoidance car systems using Machine Learning (Convolutional Neural Networks). My images are 200x100 JPG images and the output is an array of 4 element
In this repository you have an example:
https://github.com/ZFTurbo/KAGGLE_DISTRACTED_DRIVER/blob/master/run_keras_simple.py
They have different folders, in every folder there is a different class of image. They load the images using OpenCV and they build arrays that contains the class of every image.