Comparative Analysis of Pre-trained CNN Models on Retinal Diseases Classification

One method to diagnose retinal diseases is by using the Optical Coherence Tomography (OCT) scans. Annually, it is estimated that around ۳۰ million OCT scans are performed worldwide. However, the process of analyzing and diagnosing OCT scan results by an ophthalmologist requires a long time so machine learning, especially deep learning, can be utilized to shorten the diagnosis process and speed up the treatment process. In this study, several pre-trained deep learning models are compared, including EfficientNet-B۰, ResNet-۵۰V۲, Inception-V۳, and DenseNet-۱۶۹. These models will be fine-tuned and trained with a dataset containing OCT scanned images to classify four retinal conditions, namely Choroidal Neovascularization (CNV), Diabetic Macular Edema (DME), Drusen, and Normal. The models that have been trained are then tested to classify the test set and the results are evaluated using a confusion matrix in terms of accuracy, recall, precision, and F۱-score. The results show that the model with the best classification results in the batch size of ۳۲ scenario is the ResNet-۵۰V۲ model with an accuracy value of ۹۸.۲۴%, precision of ۹۸.۲۵%, recall of ۹۸.۲۴%, and F۱-score of ۹۸.۲۴%. While for the batch size of ۶۴, the EfficientNet-B۰ model is the model with the best classification results with an accuracy value of ۹۶.۵۹%, precision of ۹۶.۸۴%, recall of ۹۶.۵۹%, and F۱-score of ۹۶.۵۹%.

کلیدواژه ها:

Classification ، Comparative analysis ، Deep learning ، Optical coherence tomography ، Pre-trained model ، Retinal.Introduction.