A Highly Accurate and Robust Retinal Vessel Segmentation Algorithm
MetadataShow full item record
Retinal vessel segmentation is beneficial for eye surgery and detection of diabetic retinopathy. Mathematical morphology, or top-hat reconstruction can keep desired vessel structures. Gaussian mixture model is used here to build classification model and generate binary images of retinal vessels. After experimental testing, this work achieves the best vessel tracking ability and robustness performance.
DegreeMaster of Science (M.Sc.)
DepartmentElectrical and Computer Engineering
CommitteeKo, Seok-Bum; Karki, Rajesh; Chen, Li; Chang, Gap Soo
Copyright DateOctober 2015
Retinal segmentation, mathematical morphology, Gaussian mixture model