Emotify is a unique "Music Recommendation System” that seamlessly integrates emotion detection through facial recognition. The model used in this project achieves an accuracy of approximately 72% on the provided dataset.
Click on the image to see the demo video:
The system adapts music recommendations from Spotify based on users’ emotions, identified from 7 distinctive moods:
- Angry
- Disgust
- Fear
- Happy
- Neutral
- Sad
- Surprise
- Facial Emotion Detection: Utilizes OpenCV and Keras for real-time emotion detection through webcam input.
- Integration with Spotify: Seamlessly integrates with Spotify API to play songs based on detected emotions.
- Frontend: React app with a 'Try Emotify' button that initiates the emotion detection process.
Emotify_ANKH-master: Contains the React frontend code.emotionDetection: Backend 1 responsible for facial emotion detection.songRecommender: Backend 2 that interacts with Spotify for song recommendations.certificate: Proof of participation in the Arithemania hackathon.prototype: Demo video demonstrating the model..cache,.git,node_modules,package,package-lock: Standard project files and dependencies.me: Image demonstrating the project to judges.venue: Picture of the first hackathon venue.new.txt: Backend-related file.
-
Clone the repository:
git clone https://github.com/urvashii-b/Emotify-Arithemania.git -
Navigate to the respective folders for frontend and backend setups.
-
Follow the README files in each folder for specific setup instructions.
If you want to train your own emotion recognition model using a different dataset or architecture, you can modify the main.py script and use the dataset of your choice. Be sure to update the model.h5 file with your new model.