Detection of Diabetic Retinopathy from Fundus Images through Local Binary Patterns and Artificial Neural Network
Anila V M1, Seena Thomas2
1Anila V M, M.Tech Scholar, Department of Computer Science and Engineering, LBS Institute of Technology for Women, Thiruvananthapuram, India.
2Seena Thomas, Assistant Professor, Department of Computer Science and Engineering, LBS Institute of Technology for Women, Thiruvananthapuram, India.
Manuscript received on 10 October 2016 | Revised Manuscript received on 18 October 2016 | Manuscript Published on 30 October 2016 | PP: 5-9 | Volume-6 Issue-1, October 2016 | Retrieval Number: A4728106116/16©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: Diabetic retinopathy (DR) is one of the most frequent cause of blindness and vision loss in diabetic patients. The diabetic retinopathy is detected earlier, the better the chance that it can be effectively treated and further vision loss prevented. This condition increases the importance of automated detection of the disease. This work focuses on distinguishing between diabetic retinopathy (DR) and normal fundus images by analyzing the texture of the retina background. Local Binary Patterns (LBP) are used as texture descriptors. They are the powerful grey-scale texture descriptors that is commonly used because of its computation simplicity. Local Binary Pattern is based on looking at the local variations around each pixel, and assigning labels to different local patterns and the labels are evaluated and used in the classification stage. Probabilistic Neural Network (PNN) is the classifier that is used for the classification of abnormal and healthy images. This work suggest that LBP is a robust texture descriptor for retinal images and the proposed method analyzing the retina background directly and avoiding difficult lesion segmentation such as exudates, microaneurysms etc. can be useful for diagnostic aid.
Keywords: Diabetic Retinopathy, Local Binary Patterns, Probabilistic Neural Network, Fundus Images.
Scope of the Article: Neural Network