Eye Tracking Blink Validation for Stable VR Gaze Signals

The patent introduces a quality indicator as an intermediary validation mechanism between raw blink detection and gaze tracking updates. This quality indicator assesses image quality metrics (such as pupil detection confidence, corneal reflection quality, and image signal-to-noise ratio) to determine whether a detected blink should trigger a gaze tracking update. By using this intermediary quality assessment, the system filters out spurious blink detections that occur during poor image quality conditions while maintaining sensitivity to genuine blinks during high-quality image conditions.

The patent implements a feedback mechanism where the quality indicator continuously monitors image quality conditions and provides feedback to modulate gaze tracking updates during blink events. When image quality metrics indicate poor conditions (low pupil detection confidence, poor corneal reflection quality), the feedback mechanism suppresses gaze tracking updates despite detected blinks. Conversely, when image quality is high, the system permits rapid updates. This feedback-controlled approach maintains fast response during reliable conditions while ensuring stability during uncertain conditions.

The patent employs dynamic parameter adjustment where the quality threshold is not fixed but adapts based on overall image quality conditions and historical performance metrics. The system monitors multiple image quality parameters (pupil detection confidence, corneal reflection quality, image illumination levels) and adjusts the validation threshold accordingly. During periods of consistently high image quality, the threshold can be raised to filter more aggressively. During periods of variable or poor image quality, the threshold adapts to maintain detection sensitivity. This dynamic parameter adjustment resolves the contradiction by making the threshold context-dependent rather than static.