Machine Vision-Based Driver Distraction Detection and Recognition

This paper presents a driver distraction detection method based on the improved YOLOv11n model. The proposed method incorporates several optimization modules to enhance the original YOLOv11n architecture. First, the P2, P3, and P4 detection heads are replaced with RSCD detection heads, and the P5 de...

Full description

Saved in:
Bibliographic Details
Published in:2025 5th International Conference on Sensors and Information Technology pp. 727 - 730
Main Authors: Xie, Xiao, Zhang, Hong
Format: Conference Proceeding
Language:English
Published: IEEE 21.03.2025
Subjects:
Accuracy
Attention mechanism
Attention mechanisms
Computational modeling
Deep learning
Driver behavior recognition
Load modeling
Magnetic heads
Object detection
Object detection algorithms
Road safety
Sensor systems
Sensors
Vehicles
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper presents a driver distraction detection method based on the improved YOLOv11n model. The proposed method incorporates several optimization modules to enhance the original YOLOv11n architecture. First, the P2, P3, and P4 detection heads are replaced with RSCD detection heads, and the P5 detection head is reduced, using shared convolution to significantly decrease the model's parameter count, making it more lightweight. Additionally, the CAA attention mechanism is introduced to improve the model's global feature extraction capabilities, especially in complex scenarios and small object detection tasks, where it performs exceptionally well. Furthermore, the original loss function is replaced with EMA-SlideLoss, leading to further improvements in detection accuracy. Experimental results show that the modified YOLOv11n model achieves a 2.1% increase in accuracy, with notable improvements in recall and mAP, while maintaining a compact model size and parameter count. These findings demonstrate the effectiveness of the proposed improvements in enhancing both detection accuracy and model efficiency. Overall, the improved YOLOv11n model provides an efficient and accurate solution for detecting driver distraction, with significant implications for road safety and traffic regulation.
DOI:10.1109/ICSI64877.2025.11009365