Integrating Machine Learning for Early Diagnosis and Prognostic Assessment of Retinopathy of Premature Babies

Retinopathy of Prematurity (ROP) is a significant threat to premature babies’ vision, needing early diagnosis and prognostic assessment. Existing systems frequently need more subjectivity and high accuracy, making prompt intervention difficult. A machine learning (ML)-based method for ROP diagnosis and prognosis is provided to address these challenges. The system uses big datasets to train ML algorithms, automating and standardizing the diagnosis process. The system enables ROP early detection and accurate prognosis by combining supervised and unsupervised learning techniques, feature selection, and dimensionality reduction. When compared to existing systems, it outperforms them in terms of accuracy (0.92vs. 0.86), sensitivity (0.89vs. 0.75), specificity (0.88vs. 0.74), and AUC-ROC (0.S5vs. 0.76). Notably, the proposed system performs well across a variety of machine learning models, including Random Forest (0.92), Gradient Boosting (0.91), and Support Vector Machine (0.S9). These results highlight the system’s potential to transform ROP care, resulting in improved visual outcomes for premature babies.