FPGA based automated glaucoma detection from retinal fundus image

Abstract

Glaucoma is an autistic eye disease that causes irreversible blindness worldwide. Every day, a huge number of people in the world lose their vision permanently due to glaucoma. In 2020, about 80 million people had glaucoma worldwide, and this number is expected to increase to over 111 million by 2040. One of the dangerous characteristics of glaucoma is that it gradually increases with no pain and no noticeable symptoms to the patients in its early stage. Therefore, glaucoma is called the silent killer of sight. It damages the optic nerve that carries information from the eye to the brain and results in permanent vision loss. Though glaucoma cannot be cured fully, it can be controlled and minimize the risk of vision loss by being detected in an early stage. About 50% of blindness, which is caused due to glaucoma, can be prevented if it is detected and diagnosed early. Therefore, early detection of glaucoma is an important challenge and it is the primary treatment for preventing vision loss. Manual analysis of glaucoma detection is time-consuming and costly. In these cases, the accuracy varies by different factors. Hence, there arises the need for an automated method that detect glaucoma and its different level accurately. The fundamental measure to detect glaucoma is the CDR (Optical Cup to Disk to ratio) of the retina and the main challenge here is to localize and segment the optical disk (OD), and optical cup (OC) efficiently from retinal fundus images. FPGA image processing performs compute-intensive video and photo processing with the use of devoted hardware that supplies low latency and excessive throughput computation. A convolutional neural network can easily classify glaucoma disease after correctly training the retinal fundus images. This research proposed an efficient method to detect glaucoma automatically and classify its different stages using a convolutional neural network. The proposed method shows good performance with 95% accuracy, 97% precision, 92% recall, and a 0.94 % F1 score. For this work, we use MATLAB and Python for Software simulation. We have also developed an FPGA-based automated method to detect glaucoma for hardware simulation, successfully identified the region of interest (ROI) from retinal fundus images. This research also explores how it outperforms existing works in the related field.