Development of a Helmet-Wearing Detection System Using Computer Vision Technology and Convolutional Neural Networks

Abstract

This research aims to develop a helmet-wearing detection system using computer vision technology and deep neural networks, employing Teachable Machine, a No-Code tool that enables developers and those without coding backgrounds to easily create and train AI models. The methodology involves collecting 220 images of motorcyclists, both wearing and not wearing helmets, and then using Convolutional Neural Networks (CNN) techniques to train the model via Teachable Machine. The experimental results show that the developed model can accurately distinguish between helmet-wearing and non-helmet-wearing riders with an Accuracy of 95%, Precision of 96%, Recall of 94%, and F1-score of 95%. These results indicate that using artificial intelligence and deep learning technology can significantly enhance the efficiency of helmet-wearing detection and help reduce road accidents effectively.