Catsonic: Cat Sound Classification Using Machine Learning Algorithms

Abstract

Catsonic is a unique project aimed at helping cat owners and veterinarians understand the sounds that cats make. Cats communicate through various sounds, and deciphering these can be challenging. In collaboration with cat owners and veterinarians, the researchers gathered a diverse set of cat sounds both online and from local communities in City of San Fernando, La Union. Cat sounds dataset forms the basis for training our machine learning model. The researchers approach involves using advanced technology called Convolutional Neural Networks (CNN) to classify and recognize different cat sounds. The researchers rigorously tested and evaluated the model to ensure accuracy and effectiveness in identifying various sounds, considering factors like sensitivity and specificity. The result of the researchers’ efforts is a user friendly mobile application. Catsonic provides a tool for cat owners and veterinarians to interpret and understand cat sounds better. The researchers also assessed the application's ease of use through the System Usability Scale (SUS), the developed application was evaluated in terms of its usability and found to be acceptable with a percentage of 80.13%. Catsonic is not just an innovative solution; Catsonic serves as a bridge for improved communication between humans and their feline friends. This project showcases the power of technology, specifically machine learning, in addressing real challenges faced by pet owners and veterinarians. This capstone project highlights the transformative power of technology, specifically machine learning, in addressing tangible challenges faced by pet owners and veterinarians. Catsonic stands as a testament to the potential of leveraging advanced technologies for the betterment of the human-animal bond, showcasing how innovation can enhance understanding and connection with beloved pets.