🧠👁️ WearEye： Eyelectronics for Strabismus Diagnosis

🔍 Overview

Strabismus affects millions of patients worldwide, yet traditional diagnostic tools are subjective and inconvenient. We propose Eyelectronics, a lightweight, imperceptible, AI-embedded wearable device that digitizes strabismus diagnosis via eyelid strain analysis.

🎯 Key Features
• Multidirectional Strain-Sensing: A 60μm ultrathin HMS array (0°/45°/90°) measures eyelid deformation caused by eye movements.
• Wireless & Comfortable: The system includes a flexible sensing patch, compact signal circuit, and Bluetooth transmission—optimized for clinical use.
• MRI-based Validation: 3D FEA confirms eyelid deformation correlates with eye position.

🧠 AI Models
• InceptionTime-Tiny: A lightweight deep learning model classifies eye movement direction with 96.6% accuracy.
• Regression MLP: Predicts eye movement angles with 1.2° mean absolute error.
• Code includes both training and evaluation scripts with sample data and pretrained models.

🧪 Clinical Validation

Our system achieves high agreement (ICC = 0.998) with gold-standard Hess screen test in real patient trials, offering a one-stop digital solution for strabismus angle measurement and EOM evaluation.

📁 Contents

📜 classification_code (train.ipynb)
📜 regression_code (train_regression.ipynb)
📜 data (X_test.npy, y_test.npy)
📜 README.md

🧬 Authors

Yong Yang, Xin Liu, Jiankai Tang, Yihao Chen, Xue Feng, et al.