*This project has been submitted for following categories: Microsoft Challenge, Best Native App, Best New Comers Prize
On daily basis and especially on the hackathons we over consume sugary drinks and snacks, so we decided to monitor our food consumption.
What it does
The application scans drinks and instantly displays their nutritional values, then asks users if they want to add the item to a consumed drinks list. The list contains all the items that were consumed and counts intake of all the nutrition. (Carbohydrates, Sugars, Fat, Protein, Salt). The app compares consumed nutritional values to the recommended values and alerts users.
How we built it
For the coding part, we used Android studio, the application is written in Java and xml. We also used the Microsoft custom vision Android gitpost. For scanning drinks, we used Microsoft's custom vision service. We trained the system rigorously.
Challenges we ran into
We came across with some implementation problems, also getting accurate results was a big challenge, however, after a long time of training, the system became accurate, so at the moment it detects various objects by almost 90% accuracy.
Accomplishments that we're proud of
It wooorks :x :))
What we learned
We learnt how to implement cognitive tools in various applications, while getting more comfortable with Android app development. We also learnt how to work better as a team and work in a more professional way to ensure the best results in the fastest and efficient way possible.
What's next for SugarHoneyIceTea
We don't want to stop our project on drinks, we believe that we can achieve the same results for food and almost any product that can be consumed. As a result, it will be a system that can help humans to control their diet just by scanning everything they eat or drink. We believe it will be very helpful for people who try to balance their diet. Especially for patients with diabetes as they can easily control their sugar intake. Also, we plan to use the MyFitnessPal API (was not accessible due to approval during the hack) to help people with allergies to different components of food by checking whether the product contains anything that they should avoid. Since the model we trained still has a long way to go, we plan to further improve it. We see many other various implementations of this technology in future.