Detection of Diabetic Retinopathy using Convolutional Neural Network
Razat Agarwal1, Aditya Mahamuni2, Noopur Gautam3, Piyush Awachar4, Parth Sagar5
1Razat Agarwal, Department of Computer Engineering, Rasiklal M. Dhariwal Sinhgad School of Engineering, Pune, India.
2Aditya Mahamuni, Department of Computer Engineering, Rasiklal M. Dhariwal Sinhgad School of Engineering, Pune, India.
3Noopur Gautam, Department of Computer Engineering, Rasiklal M. Dhariwal Sinhgad School of Engineering, Pune, India.
4Piyush Awachar, Department of Computer Engineering, Rasiklal M. Dhariwal Sinhgad School of Engineering, Pune, India.
5Prof. Parth Sagar, Department of Computer Engineering, Rasiklal M. Dhariwal Sinhgad School of Engineering, Pune, India. 

Manuscript received on November 15, 2019. | Revised Manuscript received on November 23, 2019. | Manuscript published on November 30, 2019. | PP: 1957-1960 | Volume-8 Issue-4, November 2019. | Retrieval Number: C6303098319/2019©BEIESP | DOI: 10.35940/ijrte.C6303.118419

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 is a medical condition in which damage occurs to the retina due to diabetes mellitus. The diagnosis of Diabetic Retinopathy through colored fundus images stand in need of experienced clinicians to identify the presence and significance of many small features, which makes it a time consuming task. In this paper, we propose a CNN based approach to detect Diabetic Retinopathy in fundus images. Data used to train the model is prepocessed by a new segmentation technique using Gabor filters. Due to small dataset, data augmentation is done to get enough data to train the model. Our segmentation model detects intricate features in the fundus images and detect the presence of DR. A high-end Graphics Processor Unit (GPU) is used to train the model efficiently. The publicly available Kaggle Dataset is used to demonstrate impressive results, particularly for a high-level classification task. On the training dataset of 14,650 images, our proposed CNN achieves a specificity of 94% and an accuracy of 69% on 3,660 validation images.
Keywords: Augmentation, CNN, Fundus, Segmentation.
Scope of the Article: Neural Information Processing.