emagonid

how the webapp looks for the game.
different terrain based on street chosen

Inspiration

Was how we all played video games, but the purpose behind playing video games was not really justified except as enjoyment. We wanted to combine enjoyment with actual unique game dynamics and something that could also help our health.

What it does

Emagonid changes the way we play the Dino game. It not only reverses the controls but also makes the path simulated like real-world streets, and you control the Dino by jumping and squatting in real life. Based on the number of squats you do, it also calculates your average calories burnt. We can

How we built it

We built Emagonid with Vanilla JavaScript, Phaser 3, and Vite. Body tracking runs in real time on the webcam feed using MediaPipe Pose Landmarker. This is a pre-trained model that helps identify different points on the body which are then feeded to a classifier. A movement classifier detects squats and jumps. An input controller reverses those actions to drive the Dino.

Challenges we ran into

Initially, understanding how we could implement real-world human jumping and squatting to reflect the Dino's actions was a problem that we faced. Furthermore, we were thinking about how else we could make this game much more useful to an average person who uses it.

Accomplishments that we're proud of

Working prototype made.

What we learned

We learnt about new tools like Phaser and MediaPipe that made the game development process much simpler. Furthermore we understood how to make classifiers for identifying squats and jumps for real time game control. Overall, we discovered new tools and frameworks for using to make cool projects like this.

What's next for emagonid

Making a much better simulator for the dino and the street conditions. We are still unsure about this, but it would be cool to include a real world imitation of weather and street conditions on which the dino runs.