Inspiration
Our inspiration was to create a fun and innovative way to choose what your next meal will be (instead of wasting hours debating with friends).
What it does
It suggests to the user which meal at which restaurant the ai thinks is ideal for them. You can also input your information by using only your face!
How we built it
For the face movement ai, we use the opencv2 library of python to detect the face and eyes, as well as write questions and response choices on the video screen. Eyes and Face features are defined using cascade xml files. When too many eyes are detected, the topmost eyes are kept, because usually, it is features of the nose or mouth that are confused as eyes. Face inclination is then calculated from the height difference vs width difference between the eyes. To request from the API, the python requests library was used. The project uses Artificial Intelligence to suggest the meal. It takes a few pieces of information about the user and predicts what the user would eat using a Nearest Neighbor model that was trained on a randomly generated dataset. We did the UI with JavaFX and CSS. This consisted of making an iPhone application just like big delivery companies have. We also implicated a live time clock on the top right of the screen to imitate a normal iPhone 10+.
Challenges we ran into
For face recognition, seeing extra eyes where there were none was the main challenge we had to deal with. Many random bugs. The lack of an extensive database for the meal choosing ai. Make everything work together.
Accomplishments that we're proud of
We are proud of our API. We are also proud to have made a working face-detection ai. We also have a very functional UI. Finally, we are very proud of how we managed to make such a relatively big project with so many files languages, and frameworks work altogether.
What we learned
We also learned more about how facial recognition and API work. How artificial intelligence works. The software development and how to make everything work together.
What's next for Raw-Dish
Add dietary restrictions. Next is to have access to a larger database, refine the facial recognition, and finally join the extension to Radish.


Log in or sign up for Devpost to join the conversation.