Gaze Direction Identification Using a 3D Model of Eye

The idea of using eye movement tracking to facilitate navigation and control has provided new avenues of development in the field of computer vision technology. Nonetheless, the rapidity of eyeball movement requires more efficient methods of tracking them than are currently in use. In an extension to previous work, we have thus developed an effective method to track the multi directionality of an eyeball, using an extensive 3D-eyeball model, by means of utilising the corresponding features of the eye (its corners and pupil) to
accurately measure the position of the eyeball without using specified calibration. Experimental results show that the approach achieves successful eyeball position measurement with lower error rates than have been obtained with other methods.