What it does
Our application provides users a curated list of songs based on how they seem to be feeling from facial data. It hopes to give the user the best songs to fit the perfect vibe.
How we built it
We used Hume's API to decipher mood based on facial data. From there we fed that mood into Chroma's API which would then find the top 10 songs with the highest similarity score to that mood. The songs were generated from Spotify's API. From there, we would display the songs as well as their embedded youtube videos using Flask.
Accomplishments that we're proud of
I'm proud that we were able to create a baseline proof of concept. The application currently can take a sequence of static face images, determine the primary emotion in them, and generate a list of songs based on that emotion.
Future Plans
As of right now, the application can't take a user's livestream facial data. This is the number one goal to figure out for future endeavors. Additionally, I would like to have the song recommender take in a bunch of mood inputs from the facial data and formulate a song rec based on a weighted combination of said moods.
Log in or sign up for Devpost to join the conversation.