Author: Danilo Andrade de Jesus, Wrocław University of Technology

The limbus is defined as a border or a transition zone between the optical clear cornea and the opaque sclera [1,2]. Furthermore, it is a niche for stem cell population that can give rise to differentiated endothelial cells [3]. These cells are important for maintaining the clear state of the cornea since its deficiency results in the inability to replace the corneal epithelium resulting in chronic inflammation, vascularization, and even blindness. Once the limbus has been predicted other parameters such as the corneal centroid, which is important for corneal transplantation [4], contact lens fitting [5] or tracking the eye movements during refractive surgery [6], can be easily achieved.

Anatomically, the limbus epithelium gradually thickens toward the sclera and its radius of curvature could change abruptly at the junction of cornea and sclera creating a shallow furrow. In order to more accurately identify the position of the corneoscleral transition than the traditional imaging techniques (so called Horizontal Visible Iris Diameter or the White-to-white corneal diameter), the Eye Surface Profiler, a newly developed instrument based on the principle of profilometry, was used. Its 3D topographical data allowed to develop two new methods for limbus demarcation.

The first method used Zernike polynomials to separately model the inner (corneal) and outer (scleral) regions of the anterior eye in a circular domain. Bhatia-Wolf polynomials were used in the second method where the 3D data was enhanced using the 2D grayscale image of the eye. The corneoscleral fitting of both methods allowed to demark the corneoscleral transition zone according to its curvature as shown on figure (1). The method based on Bhatia-Wolf polynomials showed to be more precise and robust than the method based on Zernike polynomials when the corneoscleral data is not fully accessible. Our results suggest that these methodologies can be helpful to be adopted in ophthalmic applications where high precision in outlining the limbus characteristics is required.

Figure 1 – Eye Surface Profiler image with the limbus position estimated by the method based on Zernike polynomials (a) where the grey square is enlarged in (b). The yellow and white lines are the extremes of the corneoscleral transition and the red line the central position of the limbus. The image (c) outlines the corneoscleral transition through the method based on Bhatia-Wolf polynomials. The central position of the limbus (red line) obtained in (c) and its circular fitting (blue line) are shown in (d).


[1] E. M. Van Buskirk, “The anatomy of the limbus.,” Eye (Lond). , vol. 3 ( Pt 2), pp. 101–8, Jan. 1989.
[2] P. Ordonez and N. Di Girolamo, “Limbal epithelial stem cells: role of the niche microenvironment.,” Stem Cells, vol. 30, no. 2, pp.100–7, Feb. 2012.
[3] S. L. McGowan, H. F. Edelhauser, R. R. Pfister, and D. R. Whikehart, “Stem cell markers in the human posterior limbus and corneal endothelium of unwounded and wounded corneas.,” Mol. Vis., vol. 13, no. May, pp. 1984–2000, Jan. 2007.
[4] A. Langenbucher, B. Seitz, M. M. Kus, E. Vilchis, and G. O. Naumann, “Graft decentration in penetrating keratoplasty: nonmechanical trephination with the excimer laser (193 nm) versus the motor trephine.,” Ophthalmic Surg. Lasers, vol. 29, no. 2, pp. 106–13, Feb. 1998.
[5] R. B. Mandell, C. S. Chiang, and S. A. Klein, “Location of the major corneal reference points.,” Optom. Vis. Sci. , vol. 72, no. 11, pp. 776–84, Nov. 1995.
[6] J. Schwiegerling and R. W. Snyder, “Eye movement during laser in situ keratomileusis.,” J. Cataract Refract. Surg., vol. 26, no. 3, pp. 345–51, Mar. 2000.