Retinal disease recognition in fundus images using deep learning architectures | |
| Author | Rohith, Padmanabhuni |
| Call Number | AIT RSPR no.IM-22-07 |
| Subject(s) | Retina--Diseases--Imaging--Data processing Retinal Diseases Optical coherence tomography Deep learning (Machine learning) |
| Note | A research submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Information Management |
| Publisher | Asian Institute of Technology |
| Abstract | Retinal diseases, such as diabetic retinopathy and age-related macular degeneration, are becoming increasingly prevalent worldwide. Early detection and diagnosis of these diseases is crucial for preventing vision loss and improving patient outcomes. However, manual analysis of fundus images (images of the back of the eye) by trained professionals is time-consuming and subject to human error. Fundus photography is a cost effective and high-quality alternative to the traditional methods of recognizing eye diseases. To address this challenge, in this study, we propose a deep learning approach for recognizing multiple retinal diseases in fundus images along with its diagnostic information. In this study we compare the performance of several popular deep learning architectures like VGGNet, Inception, ResNet, EfficientNet and MobileNet on a large dataset (ODIR-5K) of fundus images and show that our approach can accurately identify a variety of retinal diseases with high sensitivity and specificity. In this study we found out that ResNet 152 produced the highest performance among the other models. We have also built a system that uses this model to classify retinal diseases into 8 main disease classes with multiple diagnostic information as auxiliary classifier. Overall, our research presents a promising approach to automating the detection and diagnosis of retinal diseases from fundus images using deep learning algorithms. Our system has the potential to improve the efficiency and accuracy of retinal disease diagnosis, ultimately leading to better patient outcomes. |
| Year | 2022 |
| Type | Research Study Project Report (RSPR) |
| School | School of Engineering and Technology |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Information Management (IM) |
| Chairperson(s) | Attaphongse Taparugssanagorn |
| Examination Committee(s) | Vatcharapon Esichaikul;Chaklam Silpasuwanchai |
| Scholarship Donor(s) | AIT Partial Scholarship |
| Degree | Research studies project report (M. Eng.) - Asian Institute of Technology, 2022 |