-
Notifications
You must be signed in to change notification settings - Fork 0
Home
Welcome to the FeedGoo wiki!
Next on my roadmap is to implement a machine learning image recognition classifier that accurately determines how much food is left in the bowl (0%, 25%, 50%, 100%) and based on the last time the cat was fed (within a 12 hr span), whether to feed again and if so, how much. I plan to use scipy, numpy, and scikitlearn.
Shout-out to the super easy to use Twitter API Python library Twython, which I use to tweet status updates with images once food has been dispensed. Follow through its documentation to register your application with Twitter; I also found this link helpful with the details of registering my app and getting my keys with Twitter, of which Twython's documentation is a little light.
I leverage the PyGame library to take a picture 2 seconds after food is dispensed, and then upload this alongside the tweet for visual proof that the feeder is working. This has been implemented for the automated/scheduled feedgoo.py and the manually activated man_feedgoo.py.
- Implement an ML image recognition classifier to feed the cat.
- Integrate with IFTTT to use voice commands with smart speakers to trigger a manual feeding, or to check on the status of the food.