StacksFX

A music player that takes an out of the box approach to how you find that perfect song by detecting moods

Emotion-Based??

  • Emotions are strong feelings that grow from different circumstances and situations.
  • StacksFX harnesses those emotions and detects your mood in real-time. 
  • And from that, it recommends music to you. Getting better each time, by learning from you.

How it Works?

  • StacksFX uses the camera installed in the laptops to click over 16 frames of the user in one go.
  • These images get classified and help in detecting mood. 
  • On the basis of the mood, songs get recommended. 

Browser (Client Side)

Server API 

if classified as happy

then StacksFX recommends a song tagged "happy" using Spotify's API

If classified as sad: 

then StacksFX

recommends a song tagged "sad" using Spotify's API

Cascade Classifier  

How it Works?

  • With every recommendation, there is a feedback loop; that learns preferences song by song, constantly building a custom profile of the user.
  • Recommends better songs to the user, inching closer to that perfect song. 
My emotion is my emotion, none of your emotion !

Feedback System

Under the hood

  • StacksFX uses WebRTC to click images
  • Cascade classifier of OpenCV for emotion detection 
  • Spotify API for song recommendations and parsing song metadata
  • Django for the backend
  • CSS + Bootstrap + Js for frontend

More Features to be added

  • Wider range of emotions detected like anger, surprise. 
  • Improving the feedback system. 
  • More accessibility in the future - like chrome extensions, packages, web.
  • Refactoring code - making it reusable for others.

Thank you !!

[IIIT hackathon winner] StacksFX

By Vipul Gupta

[IIIT hackathon winner] StacksFX

A revolutionary music player that recommends songs by detecting your emotions through the camera and constantly learning from your actions through machine learning with a constant feedback loop.

  • 1,966