This invention enables real-time eye tracking for natural human-computer interaction using only two standard cameras and no special equipment. By combining a fixed wide-angle camera with an active pan-tilt-zoom camera and novel 3D eyeball calibration, the system provides accurate gaze estimation under natural head movement, offering a practical, low-cost solution for interactive systems, analytics, and adaptive interfaces.
Current eye tracking systems are often limited by restrictive user constraints such as fixed head position, wearable sensors, infrared lighting, or expensive stereo camera setups. There are limitations to real-world deployment, including inconvenience, discomfort, and expense for users and organizations. Many existing systems also struggle with accuracy under head movement, lighting variation, or occlusion. A new solution is needed that enables accurate gaze tracking with minimal hardware, robustness to uncontrolled environments, and greater user friendliness.
This invention combines a fixed wide-angle camera with an active pan-tilt-zoom (PTZ) camera that dynamically zooms in on the eyes, then employs a novel calibration to recover each user’s 3D eyeball center, radius, and fovea position. A least-squares method computes the optical axis, then rotates the fovea to obtain the visual axis. The system operates entirely in visible light without infrared, stereo cameras, or room-specific setups and accommodates natural head pose changes through a two-stage iterative calibration procedure. The lab-scale prototype system has been validated using controlled user studies and webcam-based experimental gaze tracking trials.
![]()
• Enables gaze tracking from distances of ~3-11 feet using a dual-camera architecture
• Achieves compact gaze clusters with intra-cluster variation as low as ~1.09° mean angular deviation
• Eliminates the need for wearable eye-tracking devices or infrared-based systems
• Real-time 3D eyeball estimation and eye gaze tracking
• Enables accurate PoG estimation at distances beyond the ~50 cm range
• Accommodates natural head movement in uncontrolled environments
• Security and behavioral analytics systems for monitoring visual attention and supporting intent recognition in real time
• Advertising and user engagement measurement systems for tracking attention on digital displays and interactive media
• Psychology and cognitive research systems for studying attention, decision-making, and visual behavior in natural viewing conditions
• Augmented reality and virtual reality systems for adapting interfaces and interactions
• Human-computer interaction systems for hands-free control and adaptive display technologies
• United States Patent 8,885,882 – Filed 11/10/2014, Issued 11/11/2014
• United States Patent 9,311,527 – Filed 11/10/2014, Issued 4/12/2016
• United States Patent 9,953,214 – Filed 3/9/2016, Issued 4/24/2018
• Related publication available upon request
Lab-scale prototype system validated using controlled user studies and webcam-based experimental gaze tracking trials. TRL ~4–5.
This technology is available for licensing.
Strong licensing potential in human-computer interaction, adaptive interfaces, security analytics, AR/VR, and attention-monitoring markets where low-cost, head-free gaze tracking can enable scalable commercial deployment.
Prototype performance data, user study results, and additional technical details available upon request.