International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 02 | Feb 2024
p-ISSN: 2395-0072
www.irjet.net
Diabetic Retinopathy Detection and Classification Using Deep Learning Algorithm Dr Bhramaramba Ravi1, Jai Surya Abhishek2, E Harika3, Trinadh N4 and Sai Kumar5 1Professor, Dept. of CSE, GITAM (Deemed to be University), Visakhapatnam, Andhra Pradesh, India. 2,3,4,5Student, GITAM (Deemed to be University), Visakhapatnam, Andhra Pradesh, India. ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Our project’s main objective was to serve as a
might even cause loss of vision due to the blood vessel damage inside the eye, so it is create an awareness and ensure that people are cautious about this, and we are aiming to develop a model which makes this task easy for the patients in doubt to easily asses themselves by using this architecture for premature detection of the disease condition. The major concern in identification of this disease is that there no awareness from the patient’s side regarding this condition until the vision is completely gone and the retina is completely damage’s its tissues. By this time, when the patient would have consulted the doctor, it would have always been too late to treat him/her for the cure as the condition would have worsen to a very critical state where disease treatment will be of no betterment of the condition. If a patient is aware of this condition, then they would consult the doctor immediately after noticing the initial changes in the vision, which are classified as NPDR, otherwise the condition of the eye gets worse and veins get clogged causing blockage of circulation of blood in the veins of the eyes. In this model development process, we will be using the eye’s retina images which are required as the dataset to extract necessary features using several processing methods. The dataset of images is given to the model for training, testing purposes and the final model must be capable of determining the level of the severity in the patient.
pivotal step to address the pervasive health concern that is predominant in numerous households: diabetes. Within the context of this prevalent condition, we have stressed on the critical issue of Diabetic Retinopathy, which is a common occurrence that ultimately results from diabetes. Our primary objective lies in the development and implementation of a unique and advanced predictive method that leverages sophisticated feature extraction techniques from an extensive image dataset. By employing an innovative approach rooted in Neural Network processing, we seek to establish a robust framework for accurately detecting and categorizing the severity of Diabetic Retinopathy through efficient classification mechanisms. The core essence of our project lies in its potential to transform the early prediction and classification of diabetic retinopathy severity, thereby facilitating timely and targeted medical intervention. By harnessing the power of deep learning techniques, we aspire to equip medical practitioners and healthcare professionals with an intuitive tool that not only streamlines the diagnostic process but also enhances the overall effectiveness of patient care. Our project's overarching goal is to contribute to the broader mission of improving healthcare outcomes by enabling timely interventions, thereby reducing the potential risks and complications associated with Diabetic Retinopathy. Through the integration of cutting-edge technology and medical expertise, we aim to pave the way for a more practical and responsive approach to diabetic retinopathy management.
2. Literature survey The invaluable insights gained by understanding the work of the authors Mrs. T N Anitha and Brunda, which are followed [1] The progression of technology has suggestively impacted numerous characteristics of our lives, it has also simplified complex tasks for humans and have been extensively improving efficiency. However, certain segments of society still face challenges, particularly in healthcare. Among the visual impairments associated with various diseases, Diabetic Retinopathy stands out, posing a significant threat to vision due to retinal vascular damage caused by prolonged diabetes. This condition, fueled by elevated glucose levels in the blood, often goes unnoticed until vision changes occur, primarily affecting middle-aged and older individuals. Addressing this issue requires early detection and treatment to mitigate vision loss effectively. Utilizing Fundus eye images and employing Image Processing techniques, this study aims to extract features and assess disease severity using the K-
Key Words: Capsule Network (Capsnet), Convolutional Neural Network (CNN), Diabetic Retinopathy (DR), K – Nearest Neighbors (KNN), Machine Learning (ML), NonProliferative Diabetic Retinopathy (NPDR), Neural Network (NN), Proliferative Diabetic Retinopathy (PDR), Random Forest (RF), Support Vector Machine (SVM).
1.INTRODUCTION In this project we have identified a major problem which is retinopathy it a condition present in eyes of the patient that is caused by diabetes. To better understand this medical condition, we can say that in simple terms such as, patients with diabetes could possibly tend to display abnormalities in the eyes that are affected due to the extensive fluid leakage from the eye retina which impacts and highly damages the light sensitive tissue of the eye, to avoid this condition which
© 2024, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 653