Discrete level set segmentation for pupil morphology characterization

The pupil morphological characteristics are of great interest for non invasive early diagnosis of the central nervous system response to environmental stimuli of different nature. Their evaluation in subjects suffering some typical diseases such as diabetes, Alzheimer disease, schizophrenia, drug and alcohol addiction is of concern. In this paper geometrical pupil features such as area, centroid coordinates, eccentricity, major and minor axes lengths are estimated by a procedure based on an image segmentation algorithm. It exploits the level set formulation of the related variational problem. A discrete set up of this problem is proposed: an arbitrary initial curve is evolved towards the unique optimal segmentation boundary by a difference equation. Numerical tests are performed on real pupillometry data taken in different illumination conditions showing a high degree of robustness of the shape parameters estimates.