Extraction of Capillary Non-perfusion from Fundus Fluorescein Angiogram

Capillary Non-Perfusion (CNP) is a condition in diabetic retinopathy where blood ceases to flow to certain parts of the retina, potentially leading to blindness. This paper presents a solution for automatically detecting and segmenting CNP regions from fundus fluorescein angiograms (FFAs). CNPs are modelled as valleys, and a novel technique based on extrema pyramid is presented for trough-based valley detection. The obtained valley points are used to segment the desired CNP regions by employing a variance-based region growing scheme. The proposed algorithm has been tested on 40 images and validated against expert-marked ground truth. In this paper, we present results of testing and validation of our algorithm against ground truth and compare the segmentation performance against two others methods. The performance of the proposed algorithm is presented as a receiver operating characteristic (ROC) curve. The area under this curve is 0.842 and the distance of ROC from the ideal point (0, 1) is 0.31. The proposed method for CNP segmentation was found to outperform the watershed [1] and heat-flow [2] based methods.