Automatic Eye Biometrics from Digital Images

D.R. Iskander (Australia), S. Mioschek, M. Trunk, and W. Werth (Austria)


pupil and limbus detection, palpebral aperture, optometry


Accurate estimation of two-dimensional eye characteristics is becoming increasingly important in modern optometry and ophthalmology practices. Our particular task was to provide technical support to current studies on myopia. In this study, subjects are asked to perform a reading task dur ing which digital images of their eyes are acquired. We present a novel algorithm for automatic extraction of pupil, limbus, and palpebral fissure parameters from such images. The proposed methodology is based on the anatomical ge ometry of the human eye. It utilises several customised image processing techniques that form a robust and com putationally efficient procedure. The application of the de veloped algorithm to patient diagnosis significantly reduces the burden associated with the previous manual human eye biometrics.

Important Links:

Go Back