Chillbot is a google home based application that uses tensorflow and machine learning to determine which kinds of drinks are available in a fridge.
- First step of the project was to take as many pictures of a few types of drinks (Coke/Perrier/Diet-Coke) in various orientation and lighting conditions.
- The purpose of these images were to be used with tensorflow to help classify images that will be taken when the user queries the google home for information.
- Using the camera module from the maker kit, we could take a picture inside the fridge and process the image to determine what drinks are available.
- This information is then stored to a firebase database as a simple json containing boolean values for the type of drinks we chose demo with.
- The google home was used to create a few simple actions that the user could use to determine if the type of drink they were looking for was in the fridge.