Automatic Diagnosis of Astigmatism for Pentacam Sagittal maps Sarah Ali Hasan Research Scholar, EIED, Thapar University Patiala, India Sarah.a.hasan@gmail.com Dr. Mandeep Singh Asst. Prof. EIED, Thapar University Patiala, India mdsingh@thapar.edu Abstract—Astigmatism is very common eye disorder, which coexist with other refractive errors. Two third of the population worldwide who have myopia has astigmatism as well. Many diagnostic methods to detect these visual disorders have been developed in the last 10 years. Most of them are computer based techniques for better accuracy and fast response. In this paper an image processing based method for diagnosing astigmatism and classifying its types has been presented. The proposed technique is implemented on sagittal maps from latest topographic instruments. Keywords— Astigmatism diagnosis; Corneal Topography; Shape recognition. I. INTRODUCTION During the process of normal vision; light first enters the eye through the cornea, then passes through the pupil which works as a diaphragm to control the amount of light entering to the eye. The lens cortex focuses the light on the retina at the back of the eye. Hence, we are able to see because of the ability of the front parts of the eye to refract or bend light, i.e. cornea and the natural lens [1]. Astigmatism is caused by a deformation in the eye, in which the shape of the cornea is more oval and asymmetric than the normal round shape as shown in Figure 1. This deformity causes the light to focus on in front of and/or behind the retina instead of the retina itself; causing images to be blurred or distorted. This may further trigger headache and/or eye strain. It has been established that the cornea alone is responsible of 70% of the total refractive power in the human eye [1-3]. Keeping this motivation, the present paper discusses the corneal astigmatism only. Since the symptoms of astigmatism and other refractive disorders are almost similar, therefore, early detection of astigmatism will enable ophthalmologists to decide the proper treatment (surgery or contact lens) for the patient. In this paper an algorithm for automatic detection and classification of astigmatism is proposed. The technique is based on the morphology and pattern recognition. Topographic images of astigmatism have some typical shapes, which have been matched using image processing techniques. The proposed algorithm detect the astigmatism using specific 'bowtie' shapes and further classify them into different types. The organization of the paper is as follows. The section I introduces the objective and significance of the work. Section II explains the basics of astigmatism and some related anatomy. This will provide preliminary bio-medical information to engineers and technologists. Previous related work has also been reported in this section. Section III is about methodology and the experimental work done. In the IV section, results are presented and briefly discussed. In the end, conclusion of the work is made in the section V. II. PHYSICS OF ASTIGMATISM A normal eye has a dome round shaped cornea with a uniform curvature, which means single refracting power all over its surface [2][3]. Refractive errors can be either spherical (myopia and hyperopia) or astigmatic. Fig. 1. Shape of cornea in case of Normal vision and Astigmatism.[4] 472 978-1-4799-3080-7/14/$31.00 c 2014 IEEE