FHRFormer: A Self-supervised Transformer Approach for Fetal Heart Rate Inpainting and Forecasting

2025-09-25

AI Summary

Fetal heart rate (FHR) monitoring is prone to signal loss due to sensor displacement, and conventional interpolation methods fail to preserve spectral characteristics—impeding AI-based analysis and clinical risk assessment. To address this, we propose FHRFormer, the first self-supervised Transformer model specifically designed for FHR signal restoration. It introduces a masked autoencoding architecture that jointly models temporal dynamics, spatial correlations across multi-channel recordings, and frequency-domain features, enabling both variable-length gap reconstruction and forward-looking trend prediction. Leveraging spectral attention mechanisms and pretraining on large-scale unlabeled FHR data, FHRFormer significantly outperforms linear and spline interpolation baselines. It achieves high accuracy and robustness across diverse missing-segment durations, supporting both retrospective analysis of historical records and real-time anomaly detection in wearable monitoring systems.

Technology Category

• Machine Learning: Feature Construction/Reformulation
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Planning, Routing, and Scheduling: Activity and Plan Recognition

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

Approximately 10% of newborns require assistance to initiate breathing at birth, and around 5% need ventilation support. Fetal heart rate (FHR) monitoring plays a crucial role in assessing fetal well-being during prenatal care, enabling the detection of abnormal patterns and supporting timely obstetric interventions to mitigate fetal risks during labor. Applying artificial intelligence (AI) methods to analyze large datasets of continuous FHR monitoring episodes with diverse outcomes may offer novel insights into predicting the risk of needing breathing assistance or interventions. Recent advances in wearable FHR monitors have enabled continuous fetal monitoring without compromising maternal mobility. However, sensor displacement during maternal movement, as well as changes in fetal or maternal position, often lead to signal dropouts, resulting in gaps in the recorded FHR data. Such missing data limits the extraction of meaningful insights and complicates automated (AI-based) analysis. Traditional approaches to handle missing data, such as simple interpolation techniques, often fail to preserve the spectral characteristics of the signals. In this paper, we propose a masked transformer-based autoencoder approach to reconstruct missing FHR signals by capturing both spatial and frequency components of the data. The proposed method demonstrates robustness across varying durations of missing data and can be used for signal inpainting and forecasting. The proposed approach can be applied retrospectively to research datasets to support the development of AI-based risk algorithms. In the future, the proposed method could be integrated into wearable FHR monitoring devices to achieve earlier and more robust risk detection.

Problem

• Handling missing fetal heart rate data caused by sensor displacement during monitoring
• Preserving spectral characteristics of signals when reconstructing missing FHR data
• Enabling robust AI-based analysis despite gaps in continuous fetal monitoring

Innovation

• Masked transformer autoencoder reconstructs missing FHR signals
• Captures spatial and frequency components of data
• Enables robust signal inpainting and forecasting