Identification of Congenital Heart Defects in Ultrasound Images using U-Net Segmentation

In this study, the use of a U-Net segmentation model is measured for the accurate detection of congenital heart abnormalities in newborn ultrasound pictures. The technique includes U-Net model training, sophisticated picture preprocessing, as well as exhaustive dataset gathering. As a result of the model's considerable accuracy, sensitivity, and specificity, it has the potential to become a crucial diagnostic tool. A thorough analysis found both advantages, such as reliable performance, and disadvantages, which include the requirement for a larger and more varied dataset. Enlarging the dataset, addressing equipment variability, real-time clinical application, multidisciplinary cooperation, and ethical issues, including model improvement are among the key recommendations. In order to promote newborn healthcare, future work will require ongoing model development, real-time clinical integration, and ethical adherence. The work advances the development of accurate diagnostics in newborn care.