123-nan / emotions-based-music

Emotion-based music recommendation system

This web-based app written in Python will first scan your current emotion with the help of OpenCV & then crop the image of your face from the entire frame once the cropped image is ready it will give this image to a trained MACHINE LEARNING model in order to predict the emotion of the cropped image. This will happen for 30-40 times in 2-3 seconds,. once we have a list of emotions (containing duplicate elements) with us it will first sort the list based on frequency & remove the duplicates. After performing all the above steps we will have a list containing the user's emotions in sorted order, Now we just have to iterate over the list & recommend songs based on emotions present in the list.

Installation & Run

Create a new project in Pycharm and add the above files. After that open the terminal and run the following command. This will install all the modules needed to run this app.

  pip install -r requirements.txt

To run the app, type the following command in the terminal.

  streamlit run app.py

Libraries

• Streamlit
• Opencv
• Numpy
• Pandas
• Tensorflow
• Keras