Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Maternal and Fetal Health Status Assessment by Using Machine Learning on Optical 3D Body Scans

Ruting Cheng, Yijiang Zheng, Boyuan Feng, Chuhui Qiu, Zhuoxin Long, Joaquin A. Calderon, Xiaoke Zhang, Jaclyn M. Phillips, James K. Hahn

TL;DR

The study addresses remote prenatal health monitoring by leveraging 3D body scans captured in the 18–24 week window. It introduces a novel two-stream neural network that combines a sequence-based RNN over abdominal level circumferences with a PCA-based global shape descriptor, fused with demographic data to predict adverse pregnancy outcomes and estimate fetal weight. Results show high accuracy for predicting preterm labor and GDM and precise fetal weight estimation within a 10% error margin, outperforming traditional anthropometry by roughly 18–22%. The work demonstrates the potential of telehealth-enabled, 3D-shape–driven prenatal assessment, while noting the need for larger, longitudinal datasets and expanded body-region coverage for broader generalization.

Abstract

Monitoring maternal and fetal health during pregnancy is crucial for preventing adverse outcomes. While tests such as ultrasound scans offer high accuracy, they can be costly and inconvenient. Telehealth and more accessible body shape information provide pregnant women with a convenient way to monitor their health. This study explores the potential of 3D body scan data, captured during the 18-24 gestational weeks, to predict adverse pregnancy outcomes and estimate clinical parameters. We developed a novel algorithm with two parallel streams which are used for extract body shape features: one for supervised learning to extract sequential abdominal circumference information, and another for unsupervised learning to extract global shape descriptors, alongside a branch for demographic data. Our results indicate that 3D body shape can assist in predicting preterm labor, gestational diabetes mellitus (GDM), gestational hypertension (GH), and in estimating fetal weight. Compared to other machine learning models, our algorithm achieved the best performance, with prediction accuracies exceeding 88% and fetal weight estimation accuracy of 76.74% within a 10% error margin, outperforming conventional anthropometric methods by 22.22%.

Maternal and Fetal Health Status Assessment by Using Machine Learning on Optical 3D Body Scans

TL;DR

The study addresses remote prenatal health monitoring by leveraging 3D body scans captured in the 18–24 week window. It introduces a novel two-stream neural network that combines a sequence-based RNN over abdominal level circumferences with a PCA-based global shape descriptor, fused with demographic data to predict adverse pregnancy outcomes and estimate fetal weight. Results show high accuracy for predicting preterm labor and GDM and precise fetal weight estimation within a 10% error margin, outperforming traditional anthropometry by roughly 18–22%. The work demonstrates the potential of telehealth-enabled, 3D-shape–driven prenatal assessment, while noting the need for larger, longitudinal datasets and expanded body-region coverage for broader generalization.

Abstract

Monitoring maternal and fetal health during pregnancy is crucial for preventing adverse outcomes. While tests such as ultrasound scans offer high accuracy, they can be costly and inconvenient. Telehealth and more accessible body shape information provide pregnant women with a convenient way to monitor their health. This study explores the potential of 3D body scan data, captured during the 18-24 gestational weeks, to predict adverse pregnancy outcomes and estimate clinical parameters. We developed a novel algorithm with two parallel streams which are used for extract body shape features: one for supervised learning to extract sequential abdominal circumference information, and another for unsupervised learning to extract global shape descriptors, alongside a branch for demographic data. Our results indicate that 3D body shape can assist in predicting preterm labor, gestational diabetes mellitus (GDM), gestational hypertension (GH), and in estimating fetal weight. Compared to other machine learning models, our algorithm achieved the best performance, with prediction accuracies exceeding 88% and fetal weight estimation accuracy of 76.74% within a 10% error margin, outperforming conventional anthropometric methods by 22.22%.

Paper Structure

This paper contains 17 sections, 8 equations, 7 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related work
  3. Association between anthropometric measurement and pregnancy health status
  4. Applications of 3D body shape in obstetrics field and advanced methodologies for extracting information from 3D body shape
  5. Dataset and pre-process
  6. Dataset
  7. Pre-process
  8. Method
  9. Sequence-dependent feature extraction stream
  10. Global feature extraction stream
  11. Joint processing layer
  12. Experiments
  13. Evaluation metrics
  14. Classification Performance
  15. Regression Performance
  16. ...and 2 more sections

Figures (7)

  • Figure 1: Comparison of Fit3D scanned model and Polycam scanned model captured simultaneously.
  • Figure 2: Extracting abdominal level circumference sequence.
  • Figure 3: Architecture of the proposed algorithm: Here, the 64 level circumferences and basic demographic information are used as input for the network. The two-stream body shape analysis branch, which is designed to process the level circumferences, consists of a supervised RNN and an unsupervised PCA.
  • Figure 4: Processing details of the $i$th element.
  • Figure 5: Heatmaps of average RNN hidden states when processing the 64 level circumferences for different tasks. The upper rows labeled "0" correspond to negative-class samples, and the lower rows labeled "1" correspond to positive-class samples.
  • ...and 2 more figures