Wearable Eye Tracker Xiaoyong Ye Franz Alexander Van
Wearable Eye Tracker Xiaoyong Ye Franz Alexander Van Horenbeke David Abbott
Index �Introduction �Background �Hardware �Software �System Design �Algorithm � Pupil Localization � Ellipse Fitting � Calibration � Homographic Mapping �Experimental Results �Future Work
Introduction �A complete system able to track the user’s eye and map the position of their pupil with the area at which they are looking at in the scene in front of them
Background �Wearable Eye-Tracking information �Who has done previous work �What they have used �Recent Methods used with eye tracker
Objectives �Hardware �Wearable �Low-Cost �Light and Confortable �Moveable eye-camera �Software �Real-Time �Accurate
Hardware �Head-Mounted Gear �Two Cameras: �Scene Camera �Eye Camera
Hardware Scene Camera �Captures the scene in front of the user �Fixed to the head Eye Camera �Captures the eye �With 5 DOF with respect to the head
System Design Eye Image Pupil Localization Scene Image Calibration Done? No Ellipse Fitting Ellipse Center Marker Detection Calculate Homography Mapping Yes
Pupil Localization �Automatic Threshold (Modified Otsu’s Method) �Image Morphology(Dilation, Erosion) �Connected Components Analysis(Find Pupil) �Pupil Center Estimation
Histogram of an Eye Image Background Pupil Graylevel Threshold
Pupil Localization Threshold Erosion Connect Components Pupil Detection Dilation Fill holes
Ellipse Fitting � 1. Updating the pupil Center � 2. Need 5 points for Fitting Ellipse model � 3. RANSAC to deal with noisy points
Ellipse Fitting �RANSAC method Edge Image Starburst Algorithm Feature Points RANSAC Ellipse Fitting
Calibration �Relationship between Ellipse center to Scene Image * = Scene Position Homography Pupil Center
Solving for homographies X’ = Hx � 8 degrees of freedom in 3 x 3 matrix H, so at least n points are sufficient to determine it � Set up a system of linear equations: Ah = 0 � where vector of unknowns h = [a, b, c, d, e, f, g, h]T � Need at least 8 eqs, but the more the better… � Solve for h. solve using least-squares = 8 pairs of
calibration method 1. Look at Scene Marker and Press corresponding number on keyboard, 2. Each marker press 2 to 3 times. 3. Randomly select 8 pairs of points to calculate Homography. (Repeatly) 3. Choose the best Homography matrix.
Mapping (x 1, y 1) (x 2, y 2)
Experimental Results �Frame rate 25/second �Accurate Pupil Ellipse �Mapping error is low( 13 pixels in 640*480 image)
Demo �Link �http: //www. youtube. com/watch? v=l. BXLps. XBGOA&c ontext=C 25 ea 4 ADOEgs. To. PDsk. Io 6 A 6 r. LXR 8 ey. Sva. Ef 82 q 6 h
Future Work �Hardware �Lighter cameras �Scene camera position �Software �Use corneal refletion �Try different mapping techniques
Thank you!
- Slides: 21