Manual diagnostic screening of high-dimensional retinal imagery is prone to human error and high labor costs. These innovations automate feature characterization through engineered temporal and spatial data integration to ensure diagnostic consistency.