An Effective Segmentation of Retinal Blood Vessels Using Optimized PCA and Morphological Operators
N.C.Santosh Kumar1, Ramudu Kama2, V.Tejaswini3 , Azmeera Srinivas4
1N C Santosh Kumar, Assistant Professor, Department of CSE,KITS College, Warangal , India.
2Ramudu Kama, Assistant Professor, Department of ECE, Kakatiya Institute of Technology and Science,Warangal, India.
3Azmeera srinivas, Assistant Professor, Department of ECE, Kakatiya Institute of Technology and Science,Warangal, India.
4V.Tejaswini, PG Scholar, Department of CSE, KITS College, Warangal, India.
Manuscript received on 1 August 2019. | Revised Manuscript received on 9 August 2019. | Manuscript published on 30 September 2019. | PP: 1362-1367 | Volume-8 Issue-3 September 2019 | Retrieval Number: B3385078219/19©BEIESP | DOI: 10.35940/ijrte.B3385.098319
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: The structure of blood vessels interior in the eye is a significant source of indicator for many diseases if exist in a human body. The extraction of vasculature network of blood vessels which plays a vital role in the study and diagnosis of many eye related diseases like diabetic retinopathy, glaucoma, and many cardiovascular diseases is a challenging task. In this paper, a new method to derive tree-shaped vasculature from retinal fundus RGB images is proposed. This proposed algorithm is performed in two stages: (1) Pre-processing stage involves Particle Swarm Optimization (PSO) algorithm to compute optimized image which holds the global optimal pixels of the input RGB image followed by conversion of PSO optimized image to gray image using PCA which is then contrast enhanced with CLAHE. (2) Post-processing stage is carried out working on the contrast enhanced gray image for attaining better accuracy of retinal blood vessel segmentation by using Thresholding as well as morphological operator. The performance measures of proposed method are evaluated on DRIVE and STARE databases and obtained best results with an average accuracy of 96.44% and proven to be an outstanding method compared to other existing retina vessel segmentation algorithms.
Keywords— Retinal Fundus Image, Particle Swarm Optimization, Thresholding, Morphological Operator.
Scope of the Article: Image Security