Automated Eye Disease Classification using MobileNetV3 and EfficientNetB0 Models using Transfer Learning
Shikha Prasher, Leema Nelson, S. Gomathi
Abstract
Healthcare is one of the industries where artificial intelligence has recently experienced a period of significant growth of health. Retinal imaging is used by ophthalmologists to identify a wide range of diseases of the eye, including capillary retinal disease, which comes on by diabetes and high blood pressure. The most reliable way to check for eye disorders is through routine ophthalmoscopies. Traditional diagnostic and treatment methods have been significantly impacted on quality of life by significant breakthroughs in AI in recent years, particularly in machine learning and deep learning. Artificial intelligence learning techniques have been trained and validated using several photos of eye illnesses as glaucoma, diabetic retinopathy, normal, and cataracts. For the purpose of predicting eye disorders, this study offered several transfer learning models. MobileNetV3 and EfticientNetB0 are two examples of Basic CNN deep transfer learning algorithms that were used in the study. The EfticientNetB0 model performed well and reached 94% accuracy, whereas the MobileNetV3 model achieved 73% accuracy.