An Efficient Automated System for Screening and Validation of Neovascularization caused by Diabetes in Retinal Images
|International Journal of Engineering Trends and Technology (IJETT)||
|© 2017 by IJETT Journal|
|Year of Publication : 2017|
|Authors : Mrs.W.V.Patil, Dr.Prema Daigavane
|DOI : 10.14445/22315381/IJETT-V47P272|
Mrs.W.V.Patil, Dr.Prema Daigavane "An Efficient Automated System for Screening and Validation of Neovascularization caused by Diabetes in Retinal Images", International Journal of Engineering Trends and Technology (IJETT), V47(8),437-444 May 2017. ISSN:2231-5381. www.ijettjournal.org. published by seventh sense research group
Neovascularisation (NV) is the serious Diabetic Retinopathy (DR) disease which causes disorder in the retinal blood vascularization due to the increase in diabetes ratio by growing new unwanted weak blood vessels over the optics disc(OD) or fovea region of the retina. Previously the evaluation of NV is done either in OD or in Fovea region of retina But our automated system aims in evaluating newly born vessels in both fovea and optic disc by using new methodology for screening of NV in entire retina by using SUSAN (Smallest Unvalued Segment Assimilating Nucleus) edge detector and line tracking algorithm for segmentation along with the classification using Support Vector Machine (SVM). These vessels are very thin to identify so pre-processing plays an important role which is done using unsharp masking, LOG (Laplacian of Gaussian) filter and high boost outlier filtering which gives better PSNR values than other pre-processing method.For segmentation, SUSAN edge detector detects the blood vascularisation with lower complexity and less computational time by extracting the features. Classification of NV is done using SVM which gives 100% specificity and 100% accuracy as compared to other method.Our proposed method helps in detecting NV at OD as well as in fovea along with the detection of the severity of disease.
 Goatman KA, Fleming AD, Philip S, Williams GJ, Olson JA, and Sharp PF. Detection of New Vessels on the Optic Disc Using Retinal Photographs. IEEE Trans Med Imaging 2011;Vol-3,N0.4
 Akram UM, Jamal I, Tariq A and Imtiaz J. Automated Segmentation of Blood Vessels for Detection of Proliferative Diabetic Retinopathy. IEEE-EMBS, IC on Biomedical and Health Informatics 2012.
 Welikala RA, Dehmeshki J, Hoppe A, Tah V, Mann S, Williamson TH,Barman SA. Automated detection of proliferative diabetic retinopathy using a modified line operator and dual classification. CM and prog in biomedicine;114(2014) 247–261
 Hassan SS & David A, Bong BL & Premsenthil MD. Detection of Neovascularization in Diabetic Retinopathy. J Digit Imaging (2012) ;25:437–444
 Zhang B, Zhang L, Karray F. Retinal vessel extraction by matched filter with first-order derivative of Gaussian. Computers Bio and Medicine 40 (2010); 438-45.
 Youssef D, Solouma NH. Accurate detection of blood vessels improves the detection of exudates in color fundus images. Computer methods and Prog in biomedicine ;108(2012) 1052–1061.
 Windera RJ, Morrowb PJ, McRitchiea IN, Bailie JR, Hartd PM . Algorithms for digital image processing diabetic retinopathy. Computerized Med Imaging and Graphics 2009; 608–622
 Aquino A, Gegúndez-Arias ME, and Marín DE. Detecting the Optic Disc Boundary in Digital Fundus Images Using Morphological, Edge Detection, and Feature Extraction Techniques. IEEE Med Imaging 2010 ;VOL.29, NO. 11.
 Gadriye D, Khandale G. Neural Network Based Method for the Diagnosis of Diabetic Retinopathy. 2014 Sixth IC on Computational Intelligence and Communication Networks; 978-1-4799-6929-6/14,2014,IEEE,DOI 10.1109/.225 1075 10.1109/CICN.2014.225
 Fadzil A, Hani M, Soomro TA, Faye I, Kamel N, Yahya N. De- noising Methods for Retinal Fundus Images. 2014 IEEE; 978-1-4799-4653-2/14/$31.00
 Patil WV, Daigavane PM . A Survey on the Automated Identification of Retinal Diseases Caused. by Diabeties. Inventi Rapid: Biomedical Engineering 2015;(1):1-6, 2014
Fovea, Neovascularization, Optic Disc, Pre-Processing, SUSAN (Smallest Unvalued Segment Assimilating Nucleus) edge detector, SVM (Support Vector Machine) classifier.