A Novel Method for Detection of Retinal Lesions Using Statistical Based Segmentation with Supervised Classifier
W. Jai Singh1, R. K. Kavitha2
1W. Jai Singh, Assistant Professor SRG, Department of MCA, Kumaraguru College of Technology, Coimbatore (Tamil Nadu), India.
2R. K. Kavitha Assistant Professor SRG, Department of MCA, Kumaraguru College of Technology, Coimbatore (Tamil Nadu), India.
Manuscript received on 12 December 2018 | Revised Manuscript received on 23 December 2018 | Manuscript Published on 09 January 2019 | PP: 198-200 | Volume-7 Issue-4S November 2018 | Retrieval Number: E2032017519/19©BEIESP
Open Access | Editorial and Publishing 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: Lesion from the retinal images are one among the main sources of visual deficiency. It impacts veins in the light-sensitive tissue known as retina. Various kinds of marks in Diabetic Retinopathy (DR) will represent the abnormalities in the retina. The automated lesion segmentation in retinal pictures is a vital task in computer-aided detection systems. The research article proposes a computational framework for detection of lesion in retina images. In the initial process, Gabor filtering technique is used to enhance the lesion regions. Second, the segmentation of the suspicious region is based on expectation maximization bootstrap subgroup and mathematical morphology. A hybrid feature set is selected from the suspicious region. Finally, a classification method is applied to pin-point the lesions in the suspicious region. The projected technique has been evaluated on two public databases: DRIVE and STARE. The experimental result shows the proficiency and viability of the proposed strategy, and it can possibly be utilized to analyze DR clinically.
Keywords: Lesion, Diabetic Retinopathy, Lesion, Segmentation, Feature Extraction, Classification, Computer Aided Detection.
Scope of the Article: Probabilistic Models and Methods