Inspiration
The inspiration comes from the central feature of Tinder (a popular dating app) that allows user to swipe to left or right to express the user's interest when he or she is presented with a Facebook profile picture of the other user.
What it does
TinderMuse gets the signal from a brain sensing headband—Muse, record and analysis the emotion response of the subject when he or she is presented with a picture containing a human face in randomized order. Using machine learning enable us to predict if the subject will like a picture of another human face or not.
How we built it
We built a web application where the user has to click the icons of a heart and a cross to show his or her interest toward the random people in a dataset of 10k pictures. The event of "click" triggers Node.js and Python algorithm to record both the brain response acquired by the Muse neuroheadset and the answers clicked. Machine learning is then applied using Python to establish the prediction and test its accuracy.
Challenges we ran into
Bi-directional communication between the web app and Python was not successfully implemented due to the shortage of time. The emotion responses had to be recorded manually and separately from the web app. The recording software used was MuLES, the data analysis will be made later on.
What's next for TinderMuse
The emotion responses are recorded by the software MuLES, and the data will be analyzed in the future.
Log in or sign up for Devpost to join the conversation.