Machine Learning based Audio Spectral Imaging and Genre Classification

In this report, learn how Queppelin Team has developed a Machine Learning based application to eliminates manual intervention and automates the process of, grouping songs according to genres and identify various song attributes, making the operation fast to process millions of songs.

This task involves the recommendation engine to calculate various properties of the song, labeling them, and training the ML model on the same data.

AI/ML based Audio Genre Classification Libraries

  • Librosa for music and audio analysis
  • Pandas and Numpy for data analysis and mathematical operations.
  • Scikit-learn, an ML library built on Scipy for data handling.
  • Keras from TensorFlow for creating and training the model.