Search papers, labs, and topics across Lattice.
This paper addresses the limitations of appearance-based gaze estimation models, which are typically trained for wide gaze angles and head poses, in the context of narrow-range, long-duration screen viewing where calibration drift is a significant issue. The authors benchmark existing methods and introduce a novel differential-gaze model, evaluating them on extensive long-duration facial video data with synchronized eye tracking. Their differential estimation approach, particularly when augmented with temporal context, significantly outperforms static methods and achieves superior agreement with reference fixations and saccades, indicating its suitability for this specific viewing regime.
Gaze estimation models trained for wide angles fail to capture the nuances of prolonged screen viewing, where a novel differential approach significantly improves accuracy and eye movement reconstruction.
Appearance-based gaze estimation offers a low-cost alternative to infrared eye tracking for screen-based behavioral and clinical applications, however existing models are typically developed for wide ranges of gaze angle and head pose. Prolonged screen viewing presents a distinct regime in which gaze remains near the screen center, head motion is limited, and calibration drift accumulates over time. In this work, we benchmark six published estimators along with a proposed differential-gaze model on 140 long-duration facial video recordings with synchronized eye tracking under subject-disjoint evaluation and a common budget of calibration frames. Differential estimation was the only static method that significantly improved upon a baseline predictor of each recording's mean gaze, and was further boosted by addition of temporal context. It also yielded the strongest agreement with reference fixations and saccades, demonstrating that low angular error alone could not validate eye movement reconstruction. These findings establish differential estimation as a promising foundation for reliable gaze tracking in long-duration, narrow-range settings.
