MoodSing

Inspiration

We got tired of Spotify not being able to read a room and ruin our vibe by playing a song that ruins the mood.

What it does

Using OpenCV and Deepface libraries, our webapp can read your mood based on your facial structure and will then either skip the current song or keep the vibes going.

How we built it

We built MoodSing one brick at a time. We first implemented our mood detection model from OpenCV and Deepface and tested it to determine if it was truly working. We then added compatibility with Spotify API to get music playing for people to like or dislike. We then added skip logic for negative emotions. Finally, we put everything together onto a single web page.

Languages/Frameworks: Python, Flask, Matplotlib, HTML/CSS, and JavaScript Tools: OpenCV, Deepface, Spotify API, and Bluetooth Compatibility

Challenges we ran into

Connecting the project to the Spotify API to get music playing was very tricky. We also wanted to implement an AI that would learn from the user's emotions and decide if the next song should be similar if the user enjoyed the song or something different if the user did not like the song. However, this proved to be too complex for the given time, and the LLMs we used were unable to provide a song URI without access to Spotify's database.