Hybrid PS-V Technique: A Novel Sensor Fusion Approach for Fast Mobile Eye-Tracking with Sensor-Shift Aware Correction

摘要

This paper introduces and evaluates a hybrid technique that fuses efficientlythe eye-tracking principles of photosensor oculography (PSOG) and videooculography (VOG). The main concept of this novel approach is to use a few fastand power-economic photosensors as the core mechanism for performing high speedeye-tracking, whereas in parallel, use a video sensor operating at lowsampling-rate (snapshot mode) to perform dead-reckoning error correction whensensor movements occur. In order to evaluate the proposed method, we simulatethe functional components of the technique and present our results inexperimental scenarios involving various combinations of horizontal andvertical eye and sensor movements. Our evaluation shows that the developedtechnique can be used to provide robustness to sensor shifts that otherwisecould induce error larger than 5 deg. Our analysis suggests that the techniquecan potentially enable high speed eye-tracking at low power profiles, making itsuitable to be used in emerging head-mounted devices, e.g. AR/VR headsets.