Proposed Algorithm

Begin Main { Input ϵ Retinal fundus image dataset  For {  Steps 1–10 Fundus image datasets ϵ five retinal types. Image pre-processing. Clustering-based automated region growing segmentation (CARGS). Extract texture features ϵ histogram, co-occurrence matrix, run length matrix, and wavelet. Generate fused hybrid-feature dataset. Pre-optimization (Fisher (F), probability of error (POE) plus average correlation (AC), and mutual information (MI)) feature selection technique employed on fused hybrid-feature dataset. Extract 30 pre-optimized fused hybrid-feature dataset. Post-optimization (correlation-based feature selection—CFS) feature selection technique and employed pre-optimized fused hybrid-feature dataset. Extract 13 post-optimized, fused hybrid-feature dataset.  End For } 10.ML classifiers are employed on post-optimized, fused hybrid-feature dataset. Output = DR classification results End main }