Principal Component Analysis based Colour Scheme Optimisation in Eye Fundus Images Contrast Enhancement for Detection and Evaluation of Drusen in Age Related Macular Degeneration Patients' Follow up

摘要

Efficiency of the patient status monitoring in Age Related Macular Degeneration cases, based on evaluation of morphological properties eye fundus images, can be significantly increased by specific contrast enhancement in the images. Objects of interest - drusen (focal deposits of extracellular debris located between the basal lamina of the retinal pigment epithelium and the inner collagenous layer of Bruch membrane) usually are represented by various intensity but the same unique color in the image. Construction of the optimal color scheme to increase the contrast of drusen can be realized by means of Principal Component Analysis, which transforms original RGB color representation into principal components space. The study demonstrates that proposed method can increase contrast-to-noise ratio of the drusen areas 10-fold or more.