Pets are known as a man’s best friend. 66% of U.S. households own a pet and yet many people have no way of understanding their pets emotions. Through our project, we will focus specifically on cats and understanding what emotions their meows represent.
This project mainly utilizes CNNs (Convolutional Neural Networks) to classify cat meows. The data consists of thousands of audio files of different cat meows that were scraped from across the internet (Youtube videos, home videos, etc.), all labeled by humans. We converted these audio files to log mel spectrograms so that we could visualize each meow and subsequently feed it to our different CNNs.