Real-Time Sleepiness Detection for Driver State Monitoring System

2015-12-19

AI Summary

To address the risk of traffic accidents caused by driver fatigue, this paper proposes a real-time drowsiness detection method. The approach integrates online dynamic template matching with Kalman filtering to achieve robust eye tracking under varying illumination and head pose disturbances. Eye state (open/closed) classification is performed at millisecond-level latency using Histogram of Oriented Gradients (HOG) features and a Support Vector Machine (SVM) classifier; an alarm is triggered when consecutive eye closure exceeds a predefined threshold. The key innovation lies in a novel tracking framework that synergistically combines adaptive template updating with motion prediction, significantly enhancing stability in complex in-vehicle environments. Experimental results demonstrate an average system response latency of less than 30 ms and an eye-closure classification accuracy of 94.2%, satisfying both real-time processing and reliability requirements for embedded automotive platforms.

Technology Category

• Computer Vision: Motion & Tracking
• Intelligent Robots: State Estimation
• Machine Learning: Hardware-aware ML

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Security and Privacy: Large-scale security measurements

Abstract

Driver face monitoring system can detect driver fatigue, which is an important factor in a large number of accidents, using computer vision techniques. In this paper we present a real-time technique for driver eye state detection. At first face is detected and the eyes are searched inside face region for tracking. A normalized cross correlation based online dynamic template matching technique with combination of Kalman filter tracking is proposed to track the detected eye positions in the subsequent image frames. Support vector machine with histogram of orientation gradient features is used for classification of state of the eyes as open or closed. If the eye(s) state is detected as closed for a specified amount of time the driver is considered to be sleeping and an alarm will be generated.

Problem

• Detect driver fatigue in real-time to prevent accidents
• Track eye states using dynamic template matching and Kalman filter
• Classify eye states as open or closed using SVM with HOG features

Innovation

• Dynamic template matching with Kalman filter
• HOG features for eye state classification
• Real-time eye tracking using computer vision