Harnessing Feature Extraction Techniques alongside CNN for Diabetic Retinopathy Detection
Fatima Patel1, Saniya Shinde2, Shivani Lingwat3, Komal Bavle4, Shweta Koparde5
1Fatima Patel, Student, Pimpri Chinchwad College of Engineering & Research, Pune, India.
2Saniya Shinde, Student, Pimpri Chinchwad College of Engineering & Research, Pune, India.
3Shivani Lingwat, Student, Pimpri Chinchwad College of Engineering & Research, Pune, India.
4Komal Bavale, Student, Pimpri Chinchwad College of Engineering & Research , Pune, India.
5Shweta Koparde, Assistance Prof. in Dept. of Computer, Pimpri Chinchwad College of Engineering & Research, Pune, India.
Manuscript received on May 25, 2020. | Revised Manuscript received on June 29, 2020. | Manuscript published on July 30, 2020. | PP: 422-425 | Volume-9 Issue-2, July 2020. | Retrieval Number: B3569079220/2020©BEIESP | DOI: 10.35940/ijrte.B3569.079220
Open Access | Ethics and Policies | Cite | Mendeley
© 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: Diabetes mellitus is a disorder that inhibits your body from properly using the energy from the food you consume. The blood vessels and blood are responsible for the transport of sugar. A hormone called insulin helps cells to take in sugar to be used as energy. Deficiency in insulin causes the disease of diabetes mellitus. One of the side effect of diabetes mellitus is diabetic retinopathy. Diabetic retinopathy is the medical condition that causes the principal vision or in rare cases entire vision loss. Diabetic retinopathy has frequent occurrences in people among 20 to 60 years. Addressing this problem, we have developed an application that saves time and gives the result of the stage of the disease. This research paper presents a CNN based system that classifies the patients in four classes as 0-no DR, 1-Mild DR, 2-Moderate DR, 3-Severe DR. The system takes the input as an image taken from a fundus camera. Image processing techniques and machine learning algorithms are used for feature extraction. The Automated screening of the retinal images would assist the doctors to easily identify the patient’s condition more precisely. With this we can easily distinguish between normal and abnormal images of the retina, this will reduce the number of inspections for the doctors.
Keywords: Diabetic Retinopathy, Image processing, K means clustering algorithm, Convolutional neural networks.