Current eye-based human–computer interfaces typically depend on camera systems or wet-electrode EEG/EOG headsets, both of which carry significant drawbacks. Camera-based tracking requires controlled lighting, substantial computational resources, and costly optical components, while wet-electrode systems are uncomfortable, intrusive, and impractical for long-term wear. Dry-electrode EOG offers a promising alternative but faces challenges such as low-amplitude signals, motion artifacts, and drift instability, which hinder consistent saccade detection. A lightweight, low-cost, and accurate wearable interface is needed to reliably decode eye movements for real-time interaction and neuro-health monitoring.
This invention presents a glasses-style wearable device equipped with dry electrodes embedded into each temple and an onboard processor. The system captures electrooculographic (EOG) signals and performs fully local, wavelet-based signal processing: baseline drift is corrected using wavelet transforms, noise is reduced through median filtering, and saccades are detected using Continuous Wavelet Transform (CWT) with Haar mother wavelets at a fixed scale (e.g., scale 20). These methods enable real-time classification of saccade direction, magnitude, and size (small, medium, large). Eye movements are then encoded into a multi-level radix-7 symbol language, supporting intuitive gesture-based human–computer interaction without cameras or external computation.
• Seamless integration of dry EOG electrodes into eyeglass frames for all-day comfort and natural wearability.
• On-device wavelet-based signal processing provides robust artifact rejection and accurate saccade detection.
• Camera-free, low-power architecture eliminates the need for bulky optics or intensive computation.
• Multi-level eye-movement encoding expands the control vocabulary far beyond binary left/right triggers.
• High accuracy in distinguishing saccades and blinks under dynamic, real-world conditions.
• Cost-effective and compatible with consumer smart-glasses platforms (e.g., Google Glass, Vuzix).
• Potential for continuous, unobtrusive neuro-ocular health monitoring.
• United States, 61/900,397, Provisional, 11/5/2013, Converted 9/23/2019
• United States, 14/533,617, Utility, 11/5/2014, Patented 5/1/2018, US 9,955,895
Prototype
This technology is available for licensing.
Attractive to developers of wearable electronics, assistive communication technologies, AR/VR interfaces, and neuro-monitoring systems seeking a low-power, camera-free method for accurate, real-time eye-movement control.
Information available upon request.