Detection and tracking of eyes for eye-camera control

摘要

We propose new algorithms to extract and track positions of eyes in a real-time video stream. For extraction of eye positions, we detect blinks with image subtraction technique. Eye lid regions are fairly small. We propose how to distinguish them from face movement. For eye position tracking, we use an updating template of Between-the-Eyes pattern instead of eyes themselves. Eyes are searched based on the current position of Between-the-Eyes and their geometrical relations in the previous frame. The pattern of Between-the-Eyes is easier to locate accurately than eye patterns. We implemented the system on a PC with Pentium III 866MHz. It runs at 30 frames per second. It detects eyes and tracks them robustly.