Feb Week 1 Updates

The team now has a few different algorithms to test on retinal images to obtain their mask. In the following week, we will apply these different algorithms to the retinal images, and we will impelement the 'GetAccuracy' function following the tests to determine which one produced the best results. Also this week, we spent some time researching and comparing our methods and algorithms. The article was written by Yavuz, Zafer, and Cemal Köse, journalists at US National Library of Medicine National Institutes of Health. In their method, the retinal image is filtered with a high-passed filter and then a few different fuzzy rules are applied to the output of the high-passed image to extract blood vessels. Their three-stage algorithm included high-pass filtering, major vessel segmentation, and fine vessel segmentation with Gaussian mixture model (GMM) classifier proposed by Roychowdhury et al. In their study, they proposed a novel blood vessel extraction approach based on image enhancement techniques and unsupervised clustering methods that included K-means and Fuzzy C-means methods.

Yavuz, Zafer, and Cemal Köse. “Blood Vessel Extraction in Color Retinal Fundus Images with Enhancement Filtering and Unsupervised Classification.” Journal of Healthcare Engineering, Hindawi, 3 Aug. 2017, www.ncbi.nlm.nih.gov/pmc/articles/PMC5559979/.