Pixel Prodigies CNNs Revolutionizing Retinopathy Screening in Preterm Infants

Time-consuming and unreliable manual examination techniques have long-limited retinopathy of prematurity (ROP) screening in preterm infants. The proposed system provides a ground-breaking method for automating ROP diagnosis using Convolutional Neural Networks (CNNs) from retinal pictures. It intends to overcome the limitations of the existing manual system by improving efficiency, reliability, and scalability. By utilizing CNNs, the system obtains substantial gains in sensitivity (up to 97.5%), specificity (up to 96.4%), accuracy (up to 96.3%), and AUC-ROC values (constantly at 0.98 or 0.99) compared to previous approaches. The system outperforms other CNN architectures due to an extra cautious technique that includes data collecting, CNN architecture selection, training, and validation. The results highlight CNNs’ transformative potential in ROP screening, which promises improved diagnosis accuracy and patient outcomes. By replacing manual assessment with automated CNN-based analysis, the proposed method ushers in a new era of ROP screening, established to transform clinical practice and enhance access to timely therapies for preterm infants globally.