Implementing an end-to-end image classification web application

Using all of the things that we learned about in the previous chapters of this book, we should now be able to implement a real-world web application that allows users to upload an image and then properly classify it.

Picking up a proper Keras model

We are going to use an existing, pre-trained Python Keras CNN model. Keras applications (https://keras.io/applications/) are a set of DL models that are available as part of the framework with pre-trained weights. Among those models is VGG16, a 16-layer CNN that was implemented by the Visual Geometry Group at the University of Oxford in 2014. This model is compatible with a TensorFlow backend. It has been trained on...