An Efficient Segmentation Based Classification of Diabetic Retinopathy Identification using CLAHE with ResNet Model
R. Rajkumar1, P. Dhanalakshmi2, R. Thiruvengatanadhan3
1R. Rajkumar, Assitant Professor, Department of Computer Science Engineering, Annamalai University, Annamalai Nagar, (Tamil Nadu), India.
2P. Dhanalakshmi, Professor, Department of Computer Science Engineering, Annamalai University, Annamalai Nagar, (Tamil Nadu), India.
3R. Thiruvengatanadhan, Assitant Professor, Department of Computer Science Engineering, Annamalai University, Annamalai Nagar, (Tamil Nadu), India.
Manuscript received on 28 June 2019 | Revised Manuscript received on 05 July 2019 | Manuscript published on 30 July 2019 | PP: 2277-2283 | Volume-8 Issue-9, July 2019 | Retrieval Number: I8434078919/19©BEIESP | DOI: 10.35940/ijitee.I8434.078919
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: Diabetic retinopathy (DR) is a widespread problem for diabetic patient and it has been a main reason for blindness in the active population. Several difficulties faced by diabetic patients because of DR can be eliminated by properly maintaining the blood glucose and by timely treatment. As the DR comes with different stages and varying difficulties, it is hard to DR and also it is time consuming. In this paper, we develop an automated segmentation based classification model for DR. Initially, the Contrast limited adaptive histogram equalization (CLAHE) is used for segmenting the images. Later, residual network (ResNet) is employed for classifying the images into different grades of DR. For experimental analysis, the dataset is derived from Kaggle website which is open source platform that attempts to build DR detection model. The highest classifier performance is attained by the presented model with the maximum accuracy of 83.78, sensitivity of 67.20 and specificity of 89.36 over compared models.
Keywords: Classification, DR, Segmentation, Deep Learning, Histogram
Scope of the Article: Classification