A Computer Vision Approach to Quantification of Corneal Neovascularization

Nesreen Otoum and Eran A. Edirisinghe


Contourlet transform, Contrast enhancement, Corneal neovascularization, Segmentation of blood vessels


This paper proposes a Contourlet Transform based approach to the segmentation of corneal blood vessels that is of clinical importance in the treatment of corneal neovascularization. The quantification of blood vessels provides means for monitoring the effect of any treatment process being followed. The proposed approach initially uses a semi automatic algorithm to detect the corneal area of a high quality colour image of an eye. Subsequently the difference image between the red and green colour planes is subjected to contrast adjustment followed by a novel contrast enhancement algorithm in the Contourlet Transform domain. The enhanced blood vessel images are finally thresholded to form binary images, using which quantification is carried out based on a measure defined as a ratio of pixels belonging to blood vessels within the area of the cornea. We provide experimental results based on four practical data sets obtained from patients suffering from different levels of corneal neovascularization.

Important Links:

Go Back