Deep Learning-based Retinal Image Enhancement Using Hybrid InceptResNet152V3 Model for Hypertensive Retinopathy Detection

Chronic high blood pressure that affects the retinal vascular system causes hypertensive retinopathy (HR). It is marked by exudates, microaneurysms, and bleeding. HR must be accurately identified and evaluated to facilitate timely medical intervention. But standard retinal fundus images have low contrast, noise, and bad illumination, which makes it hard to see vascular structures. This study creates a combined hybrid deep learning model to make retinal image augmentation and categorization better. InceptionResNet152V3, ResNet152V2, and InceptionV3 are all parts of the model. The fusion technique makes the retina more visible, as seen by performance parameters including PSNR (10.57), SSIM (0.411), accuracy (0.541), MSE, Entropy, Precision, Sensitivity, and F1-score. The model's diagnostic accuracy is better than that of independent architectures thanks to InceptionV3's spatial feature extraction and ResNet152V2's structural feature analysis. Good Health and Well-Being is helped by finding disorders that can damage vision early on. Industry, Innovation, and Infrastructure is helped by advanced AI-powered imaging technologies. It helps Sustainable Cities and Communities by providing diagnostic tools that are easy to use and widely used. The suggested technique makes it easier to automatically measure heart rate and makes clinical results more reliable.