Contents

Low-Cost Gaze Tracking with 3D Cameras

../../_images/gaze.png

At typical computer monitor distances, consumer-grade 3D cameras have sufficient resolution to capture the eye’s surface. We developed an algorithm to extract the iris surface from both eyes and use it to compute gaze using an adaptation of RANSAC. Initial experiments with this technique demonstrated roughly centimeter-scale gaze resolution over the surface of a computer monitor.

References

  1. Q. Qiu, Z. Chang, M. Draelos, J. Chen, A. Bronstein, and G. Sapiro. Low-cost gaze and pulse analysis using RealSense. In EAI International Conference on Wireless Mobile Communication and Healthcare (MOBIHEALTH), 276–279. Oct 2015. doi:10.4108/eai.14-10-2015.2261657.

  2. M. Draelos, Q. Qiu, A. Bronstein, and G. Sapiro. Intel RealSense = Real low cost gaze. In IEEE International Conference on Image Processing (ICIP), 2520–2524. Sept 2015. doi:10.1109/ICIP.2015.7351256.

OCT Visualization in Virtual Reality Small Robot Navigation and Obstacle Avoidance