Eye and Robust Iris Center Tracking with Presence of Facial Occlusions

Abstractin present era, Eye tracking is an evolving research area in the field of computer vision. For video based eye tracking system, there are a lot of concerns which has to take into count. Different types of occlusion like occlusion for eyebrows or eyelashes are the major problem for iris detection. The intensity of iris, eyelashes and eyebrows are almost similar. Hence, it is difficult to differentiate between them. In our paper, we proposed a method which will resolve the occlusion of eyelashes and eyebrows for combined eye region and iris tracking. We approached in a procedural way. Two modules are used. They are: rough eye region detection and iris center detection. In the first module, the face region is extracted from the image frame based on haar like feature. Eye region is detected by vertical and horizontal edge information of face ROI. In second module, some of processing task is done on detected rough eye region to find iris center. Here, the intensity of detected eye sub image is scaled. Otsu thresholding method is used for extracting binarized eye sub image. After this we have to select iris contour. For finding the exact iris contour, we depend on two features. Based on those two features, initially, a number of iris sub-image templates are stored. After storing them, we will use those templates for selection of exact iris contour when we need. Finally Finally, for finding iris center, ellipse fitting method will be applied in the selected iris contour. A center correction method also has been explained for correction of the inaccurate center detection due to thick patched lower intensity eyelashes with iris