IRACST - International Journal of Computer Science and Information Technology & Security (IJCSITS), Vol. 2, No. 2, April 2012 376 Eye Diseases Detection based on covariance Md Alamgir Hossain 1 , Debabrata Samanta 2 and Goutam Sanyal 3 1 Department of MCA, Calcutta Institute of Technology, West Bengal, India 2,3 Department of CSE, National Institute of Technology, Durgapur, West Bengal, India Abstract— Eye diseases are the burning diseases now-a- days. Eye diseases detection is one of the imperative problems in computer vision. It has much relevance such as face live detection and driver fatigue analysis. In this paper first, the captured images are collected from different patients and are processed for enhancement. Then image segmentation is carried out to get target regions (disease spots). Finally, analysis of the target regions (disease spots) based on covariance approach to finding the phase of the disease and then the treatment consultative module can easily be prepared on the lookout for human being. Keywords: Covariance, disease detection, morphological feature. I. INTRODUCTION According to the importance along with consideration of almost the relevant features of human face, the eyes take part in a vital role in interpreting intension and attention of human being. Detecting and tracking eyes through image sequences is one of the fundamental problems in computer vision. More and more people are anguish from some forms of eye diseases and the numbers have been rising over the years for old ages in particular. Generally the diseases infected the patients slowly and most of the cases, the patients are not aware of the diseases till their vision is seriously affected. So the earlier the doctors are able to detect the eye disease the higher the chance of the patients in preventing visual loss. But the problem with eye diseases is detection of the diseases usually is neither easy nor straightforward and detection is normally found only at a later stage. Now a day technology that was use to detect different eye diseases is by capturing optical eye images. That technology depend computers program like neural network and feature extraction algorithms to extract feature from the optical images that are necessary for abnormality classification. But because this technology uses optical imaging thus it require standard lighting condition. Hence optical images are difficult to analyze mainly due to the variation that can come with different lighting condition and reflection on the eyes. H. F. Jelinek, J. Leandro, R. M. Cesar, Jr, M. J. Cree [1] proposed that the utility of pattern analysis tools linked with a simple linear discriminate analysis that not only identifies new vessel growth in the retinal fundus but also localises the area of pathology. Huan Wang, Wynne Hsu, Kheng Guan Goh, Mong Li Lee [2] propose a novel approach that combines brightness adjustment procedure with statistical classification method and local-window-based verification strategy. James D. Weiland, Wolfgang Fink, Mark Humayun, Wentai Liu, Damien C. Rodger, Yu-Chong Ta, Mark Tarbell [3] have proposed an application specific integrated circuits (ASICs) design and they have tested to demonstrate closed loop power control and efficient micro stimulation and they novel packaging process has been developed that is capable of simultaneously forming a receiver coil, interconnects, and stimulating electrodes. Jorg Meier, Rudiger Bock, Laszlo G. Nyul, Georg Michelson [4] propose an automated system that detects glaucomatous eyes based on acquired fundus images. In contrast to other approaches they use image-based features of fundus photos that do not depend on exact measurements gained by segmentation techniques. This appearance based approach is new in the field of retina image processing. Our vision is to establish a screening system that allows fast, robust and automated detection of glaucomatous changes in the eye fundus. In this paper, we proposed a novel methodology for capturing images of different patients from several hospitals for enhancement. After that, the image segmentation is carried out to get target regions (disease spots). Finally, we analysis of the target regions (disease spots) based on covariance approach for finding the phase of the disease so that the treatment consultative module can easily be prepared on the lookout for human being.