Tuff Sk8ers

Introduction
The Team
What are the rules?
The moves
Crouch
Neutral
Jump
Left Spin
Right Spin
Testing with a visualiser
Lets rip!

Inspiration

After an impulsive purchase of a TechDeck and a sudden wave of nostalgia for the Y2K era, we wanted to bring back the childhood memories of games like Tony Hawk Pro Skater, OlliOlli and Skate with the technology of Xbox Kinect. Thus Tuff Sk8ers was born; using computer vision to turn TechDeck tricks with your fingers into the skating controls of OlliOlli.

What it does

Tuff Sk8ers is a 2D skating game controlled by real-time finger tracking. Using a webcam, players can perform gestures to control their skater, bringing a hands-on twist to the classic skating-game experience.

How we built it

We combined OpenCV and MediaPipe to capture and track hand landmarks through a webcam, then used TensorFlow and Keras to train a custom gesture-classification model. We mapped the model's predictions to keyboard inputs, allowing players to control the game through natural hand movements.

Challenges we ran into

Training a model that could reliably distinguish between gestures was a major challenge. We also had to translate noisy, real-time hand-tracking data into responsive game controls while keeping our development environments consistent across different operating systems.

Accomplishments that we're proud of

We're proud to have built a working computer-vision pipeline and trained our own gesture-classification model to control a game. Turning webcam input into playable controls was a rewarding step in bringing our idea to life. We also justified an impulsive purchase AND got steam points. A true win win.

What we learned

We gained hands-on experience with OpenCV, MediaPipe, TensorFlow, and Keras, learning how to process hand landmarks, train a computer-vision model, and integrate its predictions into a real-time application. We also learned the importance of collaboration and managing dependencies across different development environments.