SegR-Net: A deep learning framework with multi-scale feature fusion for robust retinal vessel segmentation

Jihyoung Ryu, Mobeen Ur Rehman, Imran Fareed Nizami, Kil To Chong

Abstract: Highlights•SegR-Net: Deep learning for robust retinal vascular segmentation using feature extraction, magnification, precision, and dense multiscale fusion.•Optimal and precise retinal vascular segmentation with reduced model parameters for fast and accurate results.•Accurate retinal vascular segmentation with dense multiscale feature fusion for improved accuracy and resilience.•The model utilizes module fusion to enhance segmentation performance and enable more precise delineation of retinal blood vessels.•SegR-Net surpasses state-of-the-art models, improving retinal segmentation accuracy and clinical decision-making.