Open Gaze: Open-Source Eye Tracker for Smartphone Devices Using Deep Learning

Eye tracking is vital in fields like vision research, language interpretation, and usability studies, yet traditionally focused on costly, computer-based setups. Mobile devices, despite their ubiquity, lack detailed eye motion pattern analysis. We have developed an open-source smartphone gaze tracker, inspired by Google Paper, aiming for similar precision without extra equipment. Using machine learning, our tracker competes with high-end mobile eye trackers. Leveraging the MIT Gaze Capture dataset, our research replicates significant findings in eye movement during natural image observation. We also explore smartphone gaze tracking’s potential in enhancing reading comprehension. Our results underscore eye motion analysis’s importance, enabling large-scale participation and contributing to vision research, accessibility, and healthcare applications.