Music Recommender System based on Facial Emotion Recognition
Computer Vision - Interaction - Music - 2020
CONTEXT:
This was my bachelors thesis. The motivation behind this project was to define new ways of discovering new music. I designed and developed a music recommender that uses as input the reaction of the listener to a set of songs to later give as output a recommended playlist with songs to discover.
HOW:
To do so it uses a camara to recognise facial expressions and classifies them, using a Convolutional Neural Network, into sad, neutral or happy and assigns them a score. The recommendation system is content based as the songs are tagged by their genres.
To do so it uses a camara to recognise facial expressions and classifies them, using a Convolutional Neural Network, into sad, neutral or happy and assigns them a score. The recommendation system is content based as the songs are tagged by their genres.
LEARNINGS:
From this project I discovered the potential computer vision has when it comes to new ways of interaction. It was also a great introduction to recommendation algorithms and a way of expanding my machine learning skills.
From this project I discovered the potential computer vision has when it comes to new ways of interaction. It was also a great introduction to recommendation algorithms and a way of expanding my machine learning skills.
