Diagnosing eye diseases in retinal images using deep learning methods - دانشکده فنی و مهندسی
Diagnosing eye diseases in retinal images using deep learning methods
نوع: Type: Thesis
مقطع: Segment: masters
عنوان: Title: Diagnosing eye diseases in retinal images using deep learning methods
ارائه دهنده: Provider: Hanie Zomorrodi
اساتید راهنما: Supervisors: DR. Hassan Khotanlou
اساتید مشاور: Advisory Professors:
اساتید ممتحن یا داور: Examining professors or referees: Dr. Muharram Mansoorizadeh, Dr.Razieh Torkamani
زمان و تاریخ ارائه: Time and date of presentation: 2025
مکان ارائه: Place of presentation: آمفی تئاتر مهندسی
چکیده: Abstract: Various eye diseases, including diabetic retinopathy, macular degeneration, myopia, and other eye disorders, are among the most important causes of low vision and blindness in the world. Therefore, timely and accurate diagnosis of these diseases plays a key role in preventing and averting blindness. However, today, the main method of diagnosing these diseases is manually by an ophthalmologist, which is time-consuming and significantly dependent on the experience and expertise of the physician. Therefore, the development of automated methods with high accuracy and reliability for analyzing retinal fundus images is of particular importance. In this study, a deep learning-based framework for multi-class detection of eye diseases from retinal images is presented. In this method, first, raw retinal images are subjected to a set of preprocessing and data augmentation processes to improve image quality and reduce the problem of data shortage. Then, using the Swin Transformer architecture, important and meaningful features are extracted from the images. On the other hand, in order to strengthen the model's understanding of the global relationships and dependencies between features, a Transformer Encoder has been used. Also, to improve the learning process and increase the stability of the model, we have used a combined error function technique. Finally, it should be noted that the performance of the proposed method has been investigated using different evaluation criteria and the results show a significant improvement in the accuracy and correctness of diagnosis compared to previous research. The findings of this study show that the use of advanced deep learning architectures can be used as an effective and non-invasive tool in the automatic diagnosis of eye diseases and play an important role in the development of intelligent systems to assist specialists and ophthalmologists.