Open Access

ARTICLE

Deep Learning-Based Decision Support System for Predicting Pregnancy Risk Levels through Cardiotocograph (CTG) Imaging Analysis

Ali Hasan Dakheel^1,*, Mohammed Raheem Mohammed¹, Zainab Ali Abd Alhuseen¹, Wassan Adnan Hashim^2,3

1 College of Engineering, Babylon University, Babil, 51002, Iraq
2 Computer Science Department, Bayan University, Erbil, Kurdistan, 83000, Iraq
3 Artificial Intelligence Engineering Department, College of Engineering, Al-Ayen University, Thi-Qar, 64001, Iraq

* Corresponding Author: Ali Hasan Dakheel. Email:

(This article belongs to the Special Issue: Medical Imaging Decision Support Systems Using Deep Learning and Machine Learning Algorithms)

Intelligent Automation & Soft Computing 2025, 40, 195-220. https://doi.org/10.32604/iasc.2025.061622

Received 28 November 2024; Accepted 28 January 2025; Issue published 28 February 2025

Abstract

The prediction of pregnancy-related hazards must be accurate and timely to safeguard mother and fetal health. This study aims to enhance risk prediction in pregnancy with a novel deep learning model based on a Long Short-Term Memory (LSTM) generator, designed to capture temporal relationships in cardiotocography (CTG) data. This methodology integrates CTG signals with demographic characteristics and utilizes preprocessing techniques such as noise reduction, normalization, and segmentation to create high-quality input for the model. It uses convolutional layers to extract spatial information, followed by LSTM layers to model sequences for superior predictive performance. The overall results show that the model is robust, with an accuracy of 91.5%, precision of 89.8%, recall of 90.4%, and F1-score of 90.1% that outperformed the corresponding baseline models, CNN (Convolutional Neural Network) and traditional RNN (Recurrent Neural Network), by 2.3% and 6.1%, respectively. Rather, the ability to detect pregnancy-related abnormalities has considerable therapeutic potential, with the possibility for focused treatments and individualized maternal healthcare approaches, the research team concluded.

---

Keywords

---