In a significant academic achievement, Huda Abdul Hussein Amer successfully defended her master’s thesis titled “Diagnosis of Diabetic Retinopathy based on Enhanced Convolutional Neural Network” at the College of Computer Science and Information Technology, University of Qadisiyah.

Under the supervision of Assistant Professor Dr Osama Majid Hilal, the researcher delved into the realm of detecting diabetic retinopathy through deep learning, emphasizing model performance enhancement through hyperparameter tuning. The study utilized two distinct datasets, Messidor-2 and APTOS 2019, training two well-known Convolutional Neural Network (CNN) architectures—Inception-V3 and EfficientNet B3.

The results highlighted the remarkable effectiveness of the modified EfficientNet B3 CNN in diagnosing diabetic retinopathy across the disease’s five stages, as demonstrated through a comprehensive comparison with the Inception-V3 approach.