Skip to content

Latest commit

 

History

History
26 lines (18 loc) · 1.9 KB

README.md

File metadata and controls

26 lines (18 loc) · 1.9 KB

AudioBots

Audio Explorer Challenge 2023

drawing

example workflow

This is the GitHub for the Audio Explorers Electrical Challenge 1 "Sound scene classifier for hearing aids" created by Team AudioBots

AudioBots consists of:

Dennis Chenxi Zhuang
Kasper Niklas Kjær Hansen
Kristoffer Marboe
Kristian Rhindal Møllmann

Abstract

With the need for accurate and effective methods for embedded devices such as hearing aids, interest has been sparked in the field of audio classification. This paper seeks to provide a journal of how audio classification using mel-spectrograms can be done using machine learning. Our CNN-based model proves to get an accuracy of $95.3%$ using 0.5M parameters, making it feasible to run on low-power devices. We discuss our predictions based on a confusion matrix, consider the further process and propose methods for further research.