There's a challenge for consumers ordering food from home or even in a restaurant to get a good understanding of portion sizes, ingredients that they might not be familiar with, and most importantly knowing what they will be getting and eating.
What it does
The Al Dente app by Team Flan lets consumers customize their dishes by selecting each ingredient, avoiding dishes with ingredients they are allergic to, selecting recommended dishes by popularity, and seeing the finished dish they are getting from a restaurant. We are adding an entirely new meaning to placing an order.
How we built it
Our team built the Flan app using Unity, ARCore, Cinema 4D, Sketch, Trinio (photogrammetry app), and Python for the recommendation engine.
Challenges we ran into
The photogrammetry is prone to errors requiring specific lighting and movement. Additionally, using ARCore proved to be much more difficult than expected, especially when trying to convert what was built in just Unity to the ARCore app.
Accomplishments that we're proud of
Finding, forming, and collaborating with an entirely new team. We also formed a very well-balanced team with two Unity developers, two designers, and one machine learning engineer. With this, we were able to split up tasks very efficiently and be more ambitious with the scope of our project. ARCore was new to all developers as well as photogrammetry for the designers and we successfully used both in our app.
What we learned
ARCore, photogrammetry, and collaborative filtering for the recommendation engine.
What's next for Flan
Continue to build features using ARCore and enhancing the 3D models for other restaurants. We also plan to create a shared augmented reality experience for waiters and customers. Once the customer places an order, the waiter is able to see the order. This gives the waiter an exact description of what the customer is expecting. The waiter will no longer be confused by "a little bit of cheese" since the user can put exactly how much they want before ordering. This also makes it easier for waiters to see what dishes still need to be served. For example, the color of the table or dish can slowly turn red the longer that the customer waits to be served. All of these features add on to our original goal of simplifying the ordering experience. Additionally, this app isn't limited to just restaurants. The ease of customization and to-scale visualization in augmented reality are the key benefits of our app. This can be applied to other areas such as building cars, laptops, clothes, phones, rooms, or anything where customization or personalization is involved.