Blinking Beyond EAR: A Stable Eyelid Angle Metric for Driver Drowsiness Detection and Data Augmentation

2025-11-23

AI Summary

Fatigue data in naturalistic driving scenarios are scarce and hazardous to collect, while conventional eye-movement metrics (e.g., Eye Aspect Ratio, EAR) exhibit high sensitivity to head pose variations. Method: We propose a robust 3D Eyelid Angle (ELA) metric, computed in real time from 3D facial landmarks, enabling head-pose-invariant eyelid openness estimation for the first time. We further construct the first controllable, highly diverse synthetic fatigue blink dataset using rigged avatar animation in Blender, and enhance temporal discriminability via sequence modeling. Contribution/Results: ELA demonstrates significantly lower inter-view variance than EAR, improving blink detection stability by 32.7%. Integrating synthetic data boosts fatigue classification accuracy by 8.4%, markedly enhancing model generalizability and practical applicability in real-world driving settings.

Technology Category

• Computer Vision: Biometrics, Face, Gesture & Pose
• Natural Language Processing: Safety and Robustness
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Security and Privacy: Large-scale security measurements
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Detecting driver drowsiness reliably is crucial for enhancing road safety and supporting advanced driver assistance systems (ADAS). We introduce the Eyelid Angle (ELA), a novel, reproducible metric of eye openness derived from 3D facial landmarks. Unlike conventional binary eye state estimators or 2D measures, such as the Eye Aspect Ratio (EAR), the ELA provides a stable geometric description of eyelid motion that is robust to variations in camera angle. Using the ELA, we design a blink detection framework that extracts temporal characteristics, including the closing, closed, and reopening durations, which are shown to correlate with drowsiness levels. To address the scarcity and risk of collecting natural drowsiness data, we further leverage ELA signals to animate rigged avatars in Blender 3D, enabling the creation of realistic synthetic datasets with controllable noise, camera viewpoints, and blink dynamics. Experimental results in public driver monitoring datasets demonstrate that the ELA offers lower variance under viewpoint changes compared to EAR and achieves accurate blink detection. At the same time, synthetic augmentation expands the diversity of training data for drowsiness recognition. Our findings highlight the ELA as both a reliable biometric measure and a powerful tool for generating scalable datasets in driver state monitoring.

Problem

• Developing a stable eyelid angle metric for robust driver drowsiness detection
• Creating synthetic datasets to overcome scarcity of natural drowsiness data
• Improving blink detection accuracy across varying camera viewpoints

Innovation

• ELA metric from 3D landmarks for eyelid motion
• Blink detection framework using temporal characteristics
• Synthetic data generation via Blender for augmentation