FitFrog

Collect coins and boxes for points
An enemy approaches!
Defeat it with the Droid!
Oh poop :(
Defeat it with a water bottle!
Success!
Compare your progress with friends!
Resume from where you left off, or start a new game!

Inspiration

We wanted to build something to support people in their fitness goals while making the experience fun and adventurous! The idea behind this app was to create a real-life side-scrolling game that would emulate the user’s every step which is reflected in the game. This would contribute to their overall well-being, and their progress can be shared amongst friends.

What it does

The user will be able to collect coins and complete quests by defeating monsters. This is achieved by prompting the user to take photos of real life objects that allow them to expose the monster’s weakness. For example, the fire dragon is defenseless against water based objects. Correct photo submissions of entities range from a variety of water based objects including water bottles, fire hydrants, and ice cubes. The Clarifai API will recognize these submissions and would then post a message on whether or not the user has successfully defeated the monster.

How we built it

We utilized Microsoft Azure to set up the database for our leaderboard and host the page that displays the game leaderboard. The Android app was built using Cordova and various plugins. We also used the Clarifai API to recognize objects from images taken to fight monsters. The entire application was written mainly in Javascript and a little bit of Java.

Challenges we ran into

We had a lot of difficulty trying to accurately detect the user’s movement via the accelerometer functionality on our phones. There wasn’t a reliable way to interpret the type of movement based solely on the accelerometer data due to noise generated by gravity that affected the readings.

Accomplishments that we're proud of

We've implemented an algorithm that detects user movements through the use of the accelerometer, as well as Google location services to reliably predict when the user is walking, running, or involved in some sort of physical movement. We are now also able to successfully detect any image that the user takes using their phone’s camera and match it to the keywords associated with each monster the user encounters within the game.

What we learned

We learnt how to utilize the Clarifai API to detect images and match them to specific keywords that we wanted. We also learnt how to use Microsoft Azure to host our leaderboard server.

What's next for FitFrog

We hope to integrate more active movements within the game such as jumps and leaps (pun intended). We also want to create more interesting quests that would involve different monsters to defeat. In addition to these, we hope to make the leaderboard more personalized to the user by connecting it to their Facebook page.