Title :
Iris tracking with feature free contours
Author :
Hansen, Dan Witzner ; Pece, Arthur E C
Author_Institution :
IT Univ., Copenhagen, Denmark
Abstract :
An active contour method is presented and applied to robust iris tracking. The main strength of the method is that the contour model avoids explicit "feature" detection: contours are simply assumed to remove statistical dependencies on opposite sides of the contour. The contour model is utilized in particle filtering together with the EM algorithm. The method shows robustness to light changes and camera defocusing, and makes it possible to use off-the-shelf hardware for gaze-based interaction.
Keywords :
eye; human computer interaction; image restoration; optical filters; optical tracking; EM algorithm; camera defocusing; eye tracking; feature free contours; gaze-based interaction; iris tracking; off-the-shelf hardware; particle filtering; statistical dependencies; Active contours; Bones; Cameras; Eyes; Filtering algorithms; Hardware; Humans; Iris; Light emitting diodes; Robustness;
Conference_Titel :
Analysis and Modeling of Faces and Gestures, 2003. AMFG 2003. IEEE International Workshop on
Print_ISBN :
0-7695-2010-3
DOI :
10.1109/AMFG.2003.1240845