Automated Anomaly Detection in Pregnancy: AI-Driven Ultrasound Analysis

This research explores the application of AI-driven techniques for automated anomaly detection in ultrasound scans during pregnancy. Leveraging a substantial dataset of labeled ultrasound images, this study delves into robust data collection, preprocessing methodologies, meticulous model selection, and comprehensive model training techniques. The combination of U-Net for image segmentation and ResNet for image classification was selectively chosen, followed by an in-depth analysis of their adaptability to medical imaging tasks. The study concludes by proposing a hybrid approach using both segmentation and classification for more accurate anomaly detection in ultrasound images. This study delves into a comprehensive survey of AI methodologies applied to fetal ultrasonography during the second trimester. By leveraging curated datasets and innovative AI techniques, this research navigates the evolving landscape of automated anomaly detection, aiming to enhance diagnostic capabilities and bridge accessibility gaps in prenatal care.