Audio-Classification-CNN

As a final project for the CS 254: Machine Learning course at the University of Vermont, we worked with classifying audio files based on gender and emotion.

Personal Contribution To Final Project

• Built all aspects of convolutional neural nework
  • Architecture of CNN
  • Hyperparameter tuning using KerasTuner
• Preproccessed audio WAV files to be uniform length with little silent time
• Proccessed WAV files into Mel-frequency cepstrum (MFCC) data for ML algorithims

Final Project Materials

• Paper
• Slides
• Video Presentation