GAN-Based Approach for Diabetic Retinopathy Retinal Vasculature Segmentation

Abstract

Most diabetes patients develop a condition known as diabetic retinopathy after having diabetes for a prolonged period. Due to this ailment, damaged blood vessels may occur behind the retina, which can even progress to a stage of losing vision. Hence, doctors advise diabetes patients to screen their retinas regularly. Examining the fundus for this requires a long time and there are few ophthalmologists available to check the ever-increasing number of diabetes patients. To address this issue, several computer-aided automated systems are being developed with the help of many techniques like deep learning. Extracting the retinal vasculature is a significant step that aids in developing such systems. This paper presents a GAN-based model to perform retinal vasculature segmentation. The model achieves good results on the ARIA, DRIVE, and HRF datasets.

A Hybrid GAN-BiGRU Model Enhanced by African Buffalo Optimization for Diabetic Retinopathy Detection

“…As a result, storage and shared parameters are features of an RNN. RNN also performs better when training nonlinear features from serialized data [24]. Researchers offered LSTM that has the ability to acquire the correlation knowledge among lengthy immediate sequences of data, as a solution to the issue of www.ijacsa.thesai.org RNN gradient fading while able to not understand lengthyterm historical load attributes.…”

“…These results highlight the varying performance of different classification models for diabetic retinopathy, with the "Proposed ABO Based GAN-BIGRU" model showing the highest overall performance across all metrics. Alex Net [23] Random Forest [24] VGG-NIN [13] Proposed ABO Based GAN-BIGRU www.ijacsa.thesai.org The DRIVE dataset, characterized by 10 training and 10 testing images, focuses on vessel segmentation with a resolution of 896x896 pixels. The Kaggle dataset, comprising 36.1k training and 54.6k testing images, annotates only severity levels and includes vessel segmentation at a resolution of 1281x1281 pixels.…”

Diabetic Retinopathy Lesion Segmentation Method Based on Multi-Scale Attention and Lesion Perception

“…In this paper, the segmentation performance is evaluated using Area Under the Precision-Recall Curve (AUPR), Dice coefficient (Dice), and Intersection over Union (IoU) [35,36]. Higher scores in these three metrics indicate the better segmentation capability of the proposed method.…”