Title :
A parametric approach to measuring limbus corneae from digital images
Author :
Iskander, D. Robert
Author_Institution :
Dept. of Optometry, Queensland Univ. of Technol., Brisbane, Qld., Australia
fDate :
6/1/2006 12:00:00 AM
Abstract :
It is difficult to demarcate the limbus borders in standard intensity images of the eye since the transition from the cornea (iris) to sclera is gradual. Non-parametric techniques that are currently used for this task are not sufficiently precise to be adopted in those ophthalmic applications where high precision in determining the limbus/pupil characteristics is required. This is the case, for example, in customized refractive corrections. Another aspect of limbus corneae characterization is the inability of measuring in vivo the outer outline of the limbus annulus without specialized illumination techniques. To overcome these limitations, we propose a parametric approach that utilizes a sigmoidal function fitted to radial intensity profiles of the limbus corneae. By the way of simulations, we show the superiority of the proposed parametric approach when compared to an optimized nonparametric technique. Further, we review the techniques for fitting ellipses to the estimated points of the limbus annulus and discuss clinical aspects of the proposed methodology. Initial clinical experiments showed close agreement between the proposed approach and limbus corneae estimates manually obtained by experienced clinicians.
Keywords :
biomedical optical imaging; eye; medical image processing; customized refractive corrections; digital images; eye; iris; limbus cornea characterization; ophthalmic applications; parametric approach; pupil; radial intensity profiles; sclera; sigmoidal function; Cornea; Digital images; Helium; Image edge detection; In vivo; Iris; Light sources; Lighting; Performance evaluation; Size measurement; Horizontal corneal diameter; horizontal visible iris diameter; limbus size; pupillometry; sigmoidal function; white-to-white corneal diameter; Algorithms; Artificial Intelligence; Humans; Image Enhancement; Image Interpretation, Computer-Assisted; Information Storage and Retrieval; Limbus Corneae; Ophthalmoscopy; Pattern Recognition, Automated; Reproducibility of Results; Sensitivity and Specificity; Signal Processing, Computer-Assisted;
Journal_Title :
Biomedical Engineering, IEEE Transactions on
DOI :
10.1109/TBME.2006.873387