Sustainable Deep Learning Approach using Hybrid Xception–VGG19 for Retinal Disease Detection

This study introduces a hybrid deep learning system that incorporates the Xception and VGG19 architectures to improve retinal images and help diagnose hypertensive retinopathy early, in line with the UN's Sustainable Development Goals. can be seen right away. M.Ramkumar Prabhu3 Department of Electronics and Communications Engineering Saveetha School of Engineering, Saveetha Institute of Medical and Technical Sciences (SIMATS) Chennai - 602105, India ramkumarprabhu@gmail.com K.Sathish6 Department of Electronics and Communications Engineering Saveetha School of Engineering, Saveetha Institute of Medical and Technical Sciences (SIMATS) Chennai - 602105, India skkumarsatish2024@gmail.com Finding HR as soon as possible is very important for stopping permanent visual loss and figuring out cardiovascular risk. It takes a long time and isn't very The suggested solution fixes medical imaging issues caused by poor lighting and noise. Vasculature, contrast, and texture are improved, making fundus images easier to see. The quantitative analysis demonstrates that VGG19 performs well with a PSNR of 49.45 dB, SSIM of 0.993, and MSE of 0.71. This ensures diagnosis structure and accuracy. Since ophthalmologists can see vascular issues better, they can detect problems earlier. This research improves SDG 3 (Good Health and Well-Being) by developing cost-effective, AI-driven eye screening protocols to prevent hypertension-related vision loss. The energy-efficient and scalable technique aids SDG 10 (Reduced Inequalities) by making diagnostic testing easier for poor people and SDG 9 (Industry, Innovation, and Infrastructure) by making AI technologies in healthcare systems easier to utilize. The proposed Xception-VGG19 fusion technique highlights how AI may improve retinal images and forecast diseases in medical imaging. This encourages equitable and smart healthcare.