Machine learning based segmentation of retinal vessel

• المؤلف الرئيسي: Jin, Nante
• مؤلفون آخرون: Jiang Xudong
• التنسيق: Final Year Project
• اللغة: English
• منشور في: 2019
• الموضوعات: DRNTU::Engineering::Electrical and electronic engineering
• الوصول للمادة أونلاين: http://hdl.handle.net/10356/78239
• المؤسسة: Nanyang Technological University

Segmentation of retinal vessel is very important part of diagnosis for eye diseases. Nowadays we have two methods to do the segmentation. They are unsupervised method and supervised method. The unsupervised method is using matched filter(MF),vessel tracking and deformable models. These approaches implement segmentation of retinal vessel without compare the color fusion at RGB 3 color channels. The supervised method based on green channel due to it has the highest contrast between vessel and background. And distinguish each image pixel into vessel or background after training some classifiers. The purpose of this project is to compare the accuracy of different method to segment retinal vessels and improve the classification method performance. The performance will evaluate on two public share databases DRIVE and STARE.