Convolutional neural network that detects ingredients on pizzas.
The model as an InceptionV3 architecture, which is provided by Keras.
It has been trained for around 10 epochs over 8045 labeled pizza images.
The dataset of pizza images used to train the net can be found at http://pizzagan.csail.mit.edu while labels are provided in this repository.
For more information refer to the code documentation.
Script used for the training process are model.py, preprocess.py, train.py
Script used for the user interface is ui.py
See https://share.streamlit.io/jdalcorso/pizza_detector/ui.py for the final deployment.