Health And Happiness

Login Screen - Used to keep track of previously scanned foods.
Main Menu with previously scanned items.
Response from Watson about the Banana button in the bottom right with the name of the object and the calories per serving.

Inspiration

Keeping track of what you've eaten in your diet and what you haven't is a difficult task. Having to select what is in your food makes it pretty much impossible for the casual user to get into it. So we wanted to fix that

What it does

Using a picture that you've previously taken and a picture you've taken right now, the type of food is identified using IBM Watson's Visual Recognition technology. Once the food has been identified, the data is checked with the John Snow nutrition database to get calorie data.

How We built it

We used Android Studio to build the main app interface, used Bluemix on the web to program the Watson's Visual Recognition system, and used IntelliJ to make the Java server that runs on the computer.

Challenges We ran into

First was getting the Watson project setup and identifying what we need to use. The base files within the Watson project are daunting for a beginner. We also ran into problems getting the data from Watson to match data from the John Snow Labs nutrition database. Finally, our biggest problem was getting the app to communicate with the PC server to send data to Watson as the Android API doesn't include Visual Recognition.

Accomplishments that We're proud of

We are proud that we were able to get the data from Watson to match the nutrition database. We are also proud that our backend programs to work somewhat reliably as they kept giving errord for hours on end.

What We learned

We learned some basic Javascript, Node.js and JSX to get Watson up and running. We also expanded our knowledge of the Android SDK and experimented with different objects to use in the UI. We also learned how to utilize JSON within Java to cross-reference strings between two sources.

What's next for Health And Happiness

We hope to make the UI look better by scaling the images properly. We also hope to better the connection to the backend from the device as we are using a complicated workaround that may not work in all situations. Finally, we hope to keep improving the string matching from Watson to the nutritional database as the results can be mixed at times.