Armaidi Armaidi, Maulisa Oktiana, Khairun Saddami, Roslidar Roslidar, Yunida Yunida, Abdullah Fatih Farhat
Road safety in Aceh Province has emerged as a critical issue, driven by the increasing number of vehicles and the growing complexity of traffic conditions. One crucial factor influencing road safety is the emotional state of drivers, which significantly affects their behavior and decision-making while driving. This study proposes a driver emotion detection system based on a Convolutional Neural Network (CNN) architecture and presents a comparative evaluation of two widely used object detection frameworks: YOLOv11 and Detectron2. The objective is to assess the effectiveness of both models in recognizing facial expressions, namely neutral, happy, and sad, among motorcyclists in real traffic environments. To support model training and evaluation, a dataset comprising 850 images of motorcyclists was compiled, with 248 pictures labeled as neutral, 257 as happy, 228 as sad, and 117 as multi-label. Experimental results demonstrate that the YOLOv11 model, when trained with a learning rate of 0.001, achieves optimal performance, yielding a mean Average Precision (mAP) of 0.65 at an Intersection over Union (IoU) threshold of 0.50 and 0.48 across IoU thresholds ranging from 0.50 to 0.95. The model exhibited balanced detection accuracy across all expression categories. These findings suggest that the proposed system has the potential to contribute to improved road safety by enabling real-time monitoring of emotional states that may affect motorcyclist behavior. © 2025 IEEE.
Universitas Syiah Kuala, Department of Electrical and Computer Engineering, Banda Aceh, Indonesia
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock