Localization and Segmentation of the Optic Nerve Head in Eye Fundus Images Using Pyramid Representation and Genetic Algorithms - Foundations on Natural and Artificial Computation - page 430

Localization and Segmentation of the Optic
Nerve Head in Eye Fundus Images Using
Pyramid Representation and Genetic Algorithms

José M. Molina and Enrique J. Carmona

Dpto. de Inteligencia Artificial, ETSI Informática, Universidad Nacional de
Educación a Distancia (UNED), Juan del Rosal 16, 28040, Madrid, Spain
[e-mail],
[e-mail]

Abstract. This paper proposes an automatic method to locate and seg-
ment the optic nerve head (papilla) from eye fundus color photographic
images. The method is inspired in the approach presented in [1]. Here,
we use a Gaussian pyramid representation of the input image to obtain
a subwindow centered at a point of the papillary area. Then, we apply
a Laplacian pyramid to this image subwindow and we obtain a set of
interest points (IPs) in two pyramid levels. Finally, a two-phase genetic
algorithm is used in each pyramid level to find an ellipse containing the
maximum number of IPs in an offset of its perimeter and, in this way, to
achieve a progressive solution to the ONH contour. The method is tested
in an eye fundus image database and, in relation to the method described
in [1], the proposed method provides a slightly lower performance but it
simplifies the methodology used to obtain the set of IPs and also reduces
the computational cost of the whole process.

1 Introduction

The location and segmentation of the optic nerve head (ONH) is of critical
importance in retinal image analysis. The ONH, also called optic disk or papilla,
is oval-shaped and is located in the area where all the retina nerve fibres come
together to form the start of the optic nerve that leaves the back of the eyeball.
There is an area without any nerve fibres called excavation (the centre of the
papilla) and around it another area can be found, the neuroretinal ring, whose
external perimeter delimits the papillary contour.
Support systems for the diagnosis of eye diseases based on eye fundus image
information involve semi-automatic and automatic locating of the papillary con-
tour. Specifically, the first semi-automatic strategies were based on geometric
properties of the image pixels and their intensity level values [2].
Active contour-based strategies have also shown to be useful to address this
problem. A characteristic property of this type of technique is that it is highly
dependent on a preliminary stage of contour initialization, from which the fi-
nal solution is refined. In some studies, this initialization stage has been done
manually, [3,4], and in other instances automatically [5,6].