Individuals afflicted with diabetes are at risk of developing Proliferative Diabetic Retinopathy (PDR), a retinal condition characterized by the formation of aberrant blood vessels on the retina. Prompt detection and treatment of neovascularization, a key feature of PDR, are crucial in averting potential vision impairment. Numerous studies have put forth various image distilling systems for the identification of revascularization in retinal scans. However, due to its unpredictable evolutionary pathway and diminutive acreage, accurately detecting neovascularization remains a formidable task. As a result, the application of deep learning techniques is gaining prevalence in neovascularization recognition, owing to their capacity for automated extraction of complex attributes from objects. In this scholarly manuscript, we propose a novel approach for neovascularization assessment utilizing cognitive computing. The productivity of cognitive computing strategy is assessed leveraging quadruplet pretrained deep convolutional network, namely, Alex Net, Google Net, ResNet18, and Mobile Net. The dataset utilized comprises approximately 3000 fundus images, categorized into five attributes: Mild, Moderate, NoDR, Proliferate_DR, and Severe. It is important to acknowledge that classifier performance can exhibit variability depending on the specific dataset and task at hand, thus rendering the accompanying performance table valuable for selecting an appropriate classifier based on desired levels of accuracy. This investigation contributes to the advancement of automated techniques for neovascularization detection, which can provide assistance in the assessment and care of diabetic retinopathy.