Diabetic Retinopathy Diagnosis and Categorization using Deep Learning - A Review
Autor: | Reshma R. Tharakan, Riddhi B. Shah, Dhruvi C. Jariwala, Ketki C. Pathak, Bhavya N. Patel |
---|---|
Rok vydání: | 2021 |
Předmět: |
Visual perception
genetic structures Computer science Visual impairment 02 engineering and technology Diabetic retinopathy Fundus (eye) medicine.disease eye diseases 030218 nuclear medicine & medical imaging 03 medical and health sciences 0302 clinical medicine Categorization Blurred vision Diabetes mellitus 0202 electrical engineering electronic engineering information engineering medicine Optometry 020201 artificial intelligence & image processing medicine.symptom Retinopathy |
Zdroj: | 2021 5th International Conference on Intelligent Computing and Control Systems (ICICCS). |
DOI: | 10.1109/iciccs51141.2021.9432312 |
Popis: | Diabetic Retinopathy (DR), the foremost root which leads to blindness is found among working-age adults. It is caused due to diabetes that affects human eye. When such a disease is detected, then at first it does not show any symptoms or shows only mild symptoms. Gradually, it leads to blindness. There are various symptoms of DR. They may include: fluctuating vision, blurred vision, spots floating in your vision, vision loss, empty areas in vision, impaired color vision. It is critical to detect this condition in its early stage for good diagnosis. In fact, earliest stage was unable to help in diagnosing of normal eye sight. Hence, requirement of finding a DR as early as possible increased which would prevent visual impairment for patients having elongated diabetes although one is suffering from young. Microaneurysms, exudates, neovascularization and hemorrhages all these parameters decide the acuteness of DR. DR is categorized in to five stages such as normal, mild, moderate, severe Non proliferative (NPDR) or Proliferative diabetic retinopathy patient (PDR). We aim to categorize early-stage DR for better clinical benefits with more useful means. In this project we aim to use Deep Learning algorithm. This paper comprises of analysis and evaluation of the different techniques of DR diagnosis and categorization using retinal images was regulated. Accordingly, 14 research papers were studied and analyzed to provide an examination related to extracted features, classification accuracy, and the usage of different data sets, such as Indian Diabetic Retinopathy Image Dataset (IDRiD), High-Resolution Fundus (HRF) Image Database, Kaggle dataset. IDRiD is an Indian dataset which is the first retinal database representative. It is the mixture of normal retinal dataset and diabetic retinopathy retinal eye image. All in all to show different issues and to provide results that can be helpful for researchers to obtain further research on diabetic retinopathy diagnosis and categorization. |
Databáze: | OpenAIRE |
Externí odkaz: |