| dc.description.abstract | Early fundus screening is an inexpensive and effective way to prevent blindness
caused by ophthalmic diseases in ophthalmology. Manual diagnosis is time-consuming,
error-prone, and complicated in clinical settings due to a lack of medical resources, and it
may cause the condition to worsen. Automated systems for the diagnosis of eye diseases
with the help of artificial intelligence have become a hot research area in the medical field.
Currently, most systems are designed to specifically detect eye diseases while humans can
have more than one type of retinal disease in one eye. Therefore, it is necessary to develop
an automated diagnostic model that can diagnose multiple diseases simultaneously. This
report presents a convolution neural network-based model for the diagnosis of various
retinal diseases by fundus imaging. The proposal model consists of 3 main parts: the data
preprocessing phase, which includes data normalization and enhancement, the second
phase is the modeling phase, and the last stage is the prediction phase. Recommended
CNNs include ResNet 34, ResNet 50, Efficient Net, Inception V1, Inception V3, VGG 16.
In the final model the system will give the probability of all 9 diseases in each image. I
validated the model by dividing the data into 3 sets: training set, Validation set, and testing
set, and measured performance using 4 different metrics: accuracy, recall, precision, and
area under the curve (AUC). The best results for VGG16 are 98% ,94%, 96% and 97%
respectively. | en_US |
