چکيده
The automated measurement of fetal head circumference (HC) through ultrasound imaging marks a significant advancement in prenatal care, providing essential insights into fetal growth and development. This research explores various deep learning techniques, including convolutional neural networks (CNNs), to enhance the accuracy, efficiency, and reliability of fetal HC measurement. Traditional methods of fetal HC measurement heavily rely on the expertise of sonographers, leading to potential variability and errors. The proposed automated systems leverage advanced image processing techniques and deep learning models, such as U-Net and its variants, to address these challenges by offering consistent and operator-independent measurements.
The research incorporates innovative methodologies, such as multi-task learning, segmentation strategies, and model architectures like mini-LinkNet and MiTU-Net, each fine-tuned for improved segmentation performance in noisy ultrasound environments. These methods are evaluated through metrics such as Dice similarity coefficient, Hausdorff distance, and Mean Absolute Error (MAE), demonstrating accuracy levels that meet or exceed those of manual measurements. The integration of AI in this context not only improves the precision of HC measurements but also enhances clinical efficiency, providing a reliable tool for fetal monitoring. The findings suggest that these automated techniques could become indispensable in clinical settings, reducing human error and enhancing prenatal care outcomes.
