/cameraovenspytorch

use phone or tablet to see if oven is on or off using trained pytorch neural network deployed to flask app

Primary LanguageHTMLMIT LicenseMIT

oven ready

key words:

pytorch, flask, camera, heroku, transfer learning, convolutional neural network

A simple app to check if an oven is on or off. The long term intention is to see if it would be useful for partially sighted people.

A neural network is trained using pytorch and the network is presented as a Flask app.

The neural network is specific to a single oven

I have deployed the application on heroku.

The project has been developed for me to learn some new programming skills:

  • using neural networks
  • Flask web server
  • javascript and DOM programming
  • RESTFUL web design
  • css and html5
  • deployment to a commercial web service, Heroku, AWS, Azure...

There are plenty of things still to do.

  • improve the look of the app on tablets and phone using css
  • employ speech to text so that the user can start and control the app with their voice
  • improve the model to recognise when it is looking at the oven or not and then check the on off status
  • improve the model to be general for more than one make of oven att a time
  • allow a user to load images of their own oven in on off position so that a new model can be trained and added to the system
  • add technical documentation so that others can deploy it to Heroku and train a model.
  • add documentation

oven on

oven off

Flask commands for running locally

  • FLASK_ENV=development FLASK_APP=app.py flask run

Heroku commands

  • heroku login
  • git push heroku master
  • heroku local
  • heroku logs --tail