Related Experiment Video
Updated: Jan 18, 2026
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
1.0K
Real-Time Driver Attention Detection in Complex Driving Environments via Binocular Depth Compensation and
Shuhui Zhou1, Wei Zhang2, Yulong Liu1
1CGNPC Uranium Resources Co., Ltd., Beijing 100084, China.
Sensors (Basel, Switzerland)
|September 13, 2025
Summary
This study introduces a real-time driver attention recognition system (RT-DASR) using binocular vision and AI. It achieves high accuracy and low latency, enhancing safety for drivers, especially in mining vehicles.
Area of Science:
- Computer Vision
- Artificial Intelligence
- Automotive Safety
Background:
- Driver distraction is a major cause of traffic accidents.
- Current vision-based driver attention systems lack accuracy or real-time performance.
Purpose of the Study:
- To develop a real-time driver attention state recognition method (RT-DASR).
- To improve accuracy and real-time performance in driver attention monitoring.
Main Methods:
- Binocular Vision Depth-Compensated Head Pose Estimation (BV-DHPE) using binocular cameras and YOLO11n Pose.
- Multi-source Temporal Bidirectional Long Short-Term Memory (MSTBi-LSTM) for fusing head pose, vehicle speed, and gaze semantics.
- Utilizing binocular disparity for depth compensation to enhance pose estimation accuracy.
Main Results:
- BV-DHPE reduced head pose Mean Absolute Error (MAE) by 44.7% compared to monocular methods.
- RT-DASR achieved 90.4% attention recognition accuracy with 21.5 ms latency on NVIDIA Jetson Orin.
- The system demonstrated effectiveness under challenging conditions like illumination changes and occlusion.
Conclusions:
- RT-DASR offers a high-precision, low-latency solution for driver attention recognition.
- The method enhances safety, particularly for mining vehicle drivers.
- RT-DASR is suitable for integration into advanced driver assistance systems for accident prevention.
