The image classifier being created within this project aims to sort images of "loose" grocery items (i.e. individual fruits, vegetables, bakery items, etc.) by assigning them general to specific labels (i.e. fruit, apple, royal gala).
The dataset is hosted on figshare. Code for the model is stored in the .ipynb notebook here, along with the source for the web app made with Flask as posted on Google Cloud Platform.
The Image dataset can be found here.
A tentative public version of this app is available on Google Cloud Platform.
To run offline, clone repo and install requirements.
$ git clone https://github.com/SarahEGood/grocery_classifier.git
$ pip install -r requirements.txt
From here, the app can be run locally at the root directory with Flask like so:
$ flask run
Code is made free available through MIT License. For details, see LICENSE.txt.