Detect the Coke
- The fitness industry has attracted a lot of consumers.
- Fitness watches, bands, and trackers make consumers easy to track their activity and burnt calories/energy.
- But energy intake is hard to track. Currently, users enter data manually.
- hard to estimate
- no useful functions
- ultimately discourage from using
What it does
- Allowing people to get a quick access to know the foods' nutrition.
- Providing a different view of displaying information as usual.
- Individuals who willing to spend little time for a brief nutrition overview.
- Athletes, professional trainers who track their nutrition intake seriously.
How I built it
- Big Data: gather product images, nutrition facts
- Machine Learning: calculate cascade classifiers, recognize objects
- Augmented Reality (AR): display / show nutrition facts
- Object Tracking
- React Native: deploy web app to both iOS and Android platforms
We are really collaborative!
- Wangyue Wang: Front-end design and implementation, packing web app to different platforms
- Xiaofan Ni: Object tracking and AR part
- Tianyu Wang: Building objects' cascade classifiers and gathering nutrition information
Open Source Projects We Used
Challenges I ran into
Training the object model is some sort of difficult in the 48 hours.
Accomplishments that I'm proud of
But we've done!
What's next for Nutrition Lens
- Nutrition overview
- Item size detection for better accuracy.
- Reorder like Amazon Dash (re-order)
- Advertisers (recommend competitive products)
- Sponsors (recommend sponsored products)
- Connect w/ apps (API)