Our original project was to develop a mental health app that would alert users if they found themselves in a situation that would leave them unconsciously agitated. We planned to use machine learning techniques to identify any unique electrical patterns during such a stressed state. We collected training data sets by recording the EEG's of volunteers who we exposed to two stimuli. The first was a video designed to relax them, and the second was a horror short film. Unfortunately, we were unable develop an algorithm that could reliably identify the user's emotions. However, during the course of our development, we had designed a simple game that's controls relied on using the Muse's EEG detection capabilities. After moving on from the app, we decided to further pursue the idea of designing a game. We completely fleshed out the controls and that was the origin of Snake with Yo Face.
What it does
Snake with Yo Face is a hands-free version of Snake. The player controls the first-person point-of-view game with a combination of head turns and facial gestures.
How we built it
Using the Muse headband’s built-in accelerometer, the game tracks the motion of the player’s head to direct the snake on screen. Interaction with the UI commands is handled by detecting the electrical activity of the player’s brain as they perform key facial gestures.
Challenges we ran into
We debated on how to recognize directional movement within the acceleration wave. Our original idea was to repeatedly calculate the derivative to find the first crest of the turning motion. We decided this was inefficient and instead constructed data sets for each movement. Using these sets, we calculated a firing threshold.
Accomplishments that we're proud of
We were able to bounce back from our previous failure and implement an interesting and fun to build project.
What we learned
1) Data Cleaning 2) Basic Machine Learning Concepts 3) Cross Platform Interaction with the Muse Headband
What's next for Snake with Yo Face
Revolutionizing the gaming industry