🤖 AI Summary
This work proposes a low-cost, scalable real-time gaze tracking method that operates within standard web browsers using only an ordinary webcam, enabling accessible cognitive and clinical research. The system leverages a lightweight, open-source deep learning pipeline that integrates MediaPipe for facial landmark detection with a convolutional neural network inspired by iTracker, augmented by a user-specific fine-tuning mechanism. We evaluate two training strategies—transferring a model pretrained on mobile data versus training from scratch on desktop-collected data—and find comparable performance on the MPIIFaceGaze benchmark, with fine-tuning substantially reducing prediction error. In a Dot-Probe task, the system’s left–right gaze allocation closely aligns with outputs from the commercial SeeSo SDK, demonstrating practical validity. The approach balances transparency, scalability, and ease of deployment.
📝 Abstract
We introduce EyeTheia, a lightweight and open deep learning pipeline for webcam-based gaze estimation, designed for browser-based experimental platforms and real-world cognitive and clinical research. EyeTheia enables real-time gaze tracking using only a standard laptop webcam, combining MediaPipe-based landmark extraction with a convolutional neural network inspired by iTracker and optional user-specific fine-tuning. We investigate two complementary strategies: adapting a model pretrained on mobile data and training the same architecture from scratch on a desktop-oriented dataset. Validation results on MPIIFaceGaze show comparable performance between both approaches prior to calibration, while lightweight user-specific fine-tuning consistently reduces gaze prediction error. We further evaluate EyeTheia in a realistic Dot-Probe task and compare it to the commercial webcam-based tracker SeeSo SDK. Results indicate strong agreement in left-right gaze allocation during stimulus presentation, despite higher temporal variability. Overall, EyeTheia provides a transparent and extensible solution for low-cost gaze tracking, suitable for scalable and reproducible experimental and clinical studies. The code, trained models, and experimental materials are publicly available.