Computer Aided Diagnosis of Eye Disease for Diabetic Retinopathy
| Autor: | Wan Nur Hafsha Wan Kairuddin, Audrey Huong, Wei Sheng Ng, Raja Mohd Aizat Raja Izaham, Wan Mahani Hafizah Wan Mahmud, Hong Seng Gan |
|---|---|
| Rok vydání: | 2019 |
| Předmět: | |
| Zdroj: | Journal of Physics: Conference Series. 1372:012030 |
| ISSN: | 1742-6596 1742-6588 |
| DOI: | 10.1088/1742-6596/1372/1/012030 |
| Popis: | Diabetic retinopathy (DR) is one of diabetes complication that could cause the vision loss, where it is caused by the damage of the blood vessels at the back of the eye. Due to this, regular eye check-up and timely treatment is needed. However, the lack of specialized ophthalmologists together with associated higher medical costs makes regular check-up costly. Therefore, any application of the technologies such as Computer Aided Diagnosis (CAD) system that could help in analysing DR efficiently in its early stage may help this current situation. Although CAD systems were developed before, but the graphic user interface for user is not developed for the ease of uses for everyone and not just limited to professional. So, in this study, a system is created in order to help the doctor to reduce their burden on the job daily and the false negative rate for the benefit of the patient. The input of the system is the Retinal fundus images (RFI) from STARE database, and the system was built with the ability to enhance and process the image for confirmation of DR. In addition, the system will help to extract out the important features based on Grey Level Co-Occurrence Matrix (GLCM) and classify it using artificial neural network (ANN) whether the patient is associated with the characteristics of DR. Also, the system will be easy to use to everyone as it will have its own graphic user interface to make it clear to everyone not just professionals so that the image from the RFI can be inserted and the result will come out in a short duration of time. The developed system able to achieve as high as 88% sensitivity, 80% specificity and 84% accuracy. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|