Related Experiment Video
Updated: May 20, 2026
06:34
A Single-Channel and Non-Invasive Wearable Brain-Computer Interface for Industry and Healthcare
Published on: July 7, 2023
Ultra-low-cost 3D gaze estimation: an intuitive high information throughput compliment to direct brain-machine
1Department of Bioengineering, Imperial College London, SW7 2AZ London, UK.
Journal of Neural Engineering
|July 14, 2012
Summary
This study introduces an affordable 3D eye-tracking system using video game hardware. This novel human-machine interface offers high performance for controlling devices, even for individuals with motor impairments.
Area of Science:
- Biomedical Engineering
- Human-Computer Interaction
- Neuroscience
Background:
- Eye movements correlate with motor intentions and are often preserved in patients with severe motor impairments.
- Current eye-tracking interfaces for impaired individuals suffer from poor signal interpretation and limited control flexibility.
- 3D gaze tracking offers a richer signal for human-machine interfaces (HMIs) by enabling direct environmental interaction.
Purpose of the Study:
- To develop an ultra-low-cost, high-performance binocular eye-tracker using readily available video game hardware.
- To evaluate the feasibility of 3D gaze estimation for controlling complex user interfaces and devices.
- To compare the information throughput and latency of the proposed eye-tracking system with existing brain-machine interfaces (BMIs).
Main Methods:
- Development of a head-mounted, binocular eye-tracker using mass-produced video game hardware with a material cost of approximately $30 USD.
- Implementation of 2D and 3D gaze estimation algorithms.
- Real-time control of a volumetric cursor and demonstration of closed-loop control through a novel benchmark: playing the video arcade game 'Pong'.
Main Results:
- The developed eye-tracker achieves performance comparable to commercial systems at an 800x lower cost, operating at over 120 Hz with 0.5-1 degree visual angle resolution.
- The system demonstrates an information throughput of 43 bits s(-1), exceeding that of invasive and semi-invasive BMIs by over tenfold.
- Effective real-time closed-loop control with 10 ms latency was achieved after only ten minutes of training.
Conclusions:
- 3D eye tracking using affordable, mass-produced hardware provides a viable and cost-effective alternative to current HMIs and expensive BMIs for individuals with motor deficiencies.
- The system's high performance, low latency, and ease of training enable effective real-time control of devices and complex interfaces.
- This technology has the potential to significantly enhance the quality of life and independence for individuals with severe motor impairments.
