Executive Summary : | Ultrasound imaging is a safe and cost-effective method for monitoring fetal health, particularly Fetus Birth Weight (FBW), which is a crucial parameter for antenatal care. However, there are no biomarkers or standard methods for accurate weight prediction, and high birth weight accounts for about 10% of total births. High BW contributes to morbidity in perinatal and maternal health. An AI-based support system could help clinicians choose the appropriate delivery model and ensure mental and clinical preparedness for any complications during delivery. Currently, clinicians manually measure fetus biometric parameters, such as Head Circumference, Femur Length, Abdominal Circumference, and Biparietal Diameter, from sonographic images. The FBW is computed using a regression formula, known as Hadlocks' formula. However, researchers have confirmed that the error in sonographic estimation of FBW is low for normal BW and higher for low and high BW cases. Accurate measurements are difficult to obtain from ultrasound images manually, especially for high-to-gestation-age babies, increasing the error in FBW estimation. This project aims to develop an AI model for automated measurement of FBW from ultrasound images with improved accuracy. |