Related Experiment Video
Updated: Aug 2, 2026
Instrumentation of Near-term Fetal Sheep for Multivariate Chronic Non-anesthetized Recordings
Published on: October 25, 2015
BabyNet＋＋: Fetal birth weight prediction using biometry multimodal data acquired less than 24 hours before delivery
Szymon Płotka1, Michal K Grzeszczyk2, Robert Brawura-Biskupski-Samaha3
1Sano Centre for Computational Medicine, Cracow, Poland; Informatics Institute, University of Amsterdam, Amsterdam, The Netherlands; Department of Biomedical Engineering and Physics, Amsterdam University Medical Center, Amsterdam, The Netherlands.
This study introduces an AI method for predicting fetal weight using ultrasound videos and clinical data, achieving expert-level accuracy. This approach enhances perinatal care by providing reliable birth weight estimations.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Obstetrics
Background:
- Accurate fetal weight prediction is crucial for antenatal management and delivery decisions.
- Current ultrasound methods face challenges in advanced pregnancy due to image quality issues.
- Manual fetal weight estimation carries inherent risks of error.
Purpose of the Study:
- To develop and validate a novel automated method for predicting fetal birth weight.
- To leverage fetal ultrasound video scans and clinical data for improved accuracy.
- To reduce errors associated with manual fetal weight estimation.
Main Methods:
- A Transformer-based approach combining a Residual Transformer Module and Dynamic Affine Feature Map Transform was employed.
- The method integrates 2D+t spatio-temporal features from ultrasound videos with tabular clinical data.
- Development and evaluation utilized 582 fetal ultrasound videos and clinical records from 194 patients.
Main Results:
- The proposed method demonstrated superior performance compared to state-of-the-art automated techniques.
- Fetal birth weight estimation accuracy was comparable to that of human experts.
- Automated measurements significantly reduced the risk of errors inherent in manual assessments.
Conclusions:
- The novel automated method offers a reliable tool for predicting fetal birth weight.
- This approach can aid clinicians, especially less experienced ones, in optimizing perinatal care.
- The system has the potential to improve the accuracy and efficiency of antenatal management.
