Utilization of Convolutional Neural Network (CNN) to Build a Camera-Based Personal Protective Equipment (PPE) Detection System

Occupational safety is a critical aspect in the industrial sector. One important implementation of safety standards is the use of Personal Protective Equipment (PPE). However, manual monitoring is often ineffective and requires significant human resources. This research aims to develop an automated PPE detection system using Deep Learning technology, specifically a Convolutional Neural Network (CNN) architecture, integrated with a camera. The detection model is built using the YOLOv5 algorithm, which is based on CNN principles, and trained with a dataset containing images of workers with and without PPE. The system was tested in a simulated work environment using real-time camera input. Evaluation results show that the model achieved an accuracy of XX%, indicating that the system can serve as an effective tool in supporting workplace safety and compliance.