Transabdominal Fetal Oximetry Using Mixed-PPG Deep Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing transabdominal fetal pulse oximetry (TFO) techniques face challenges in accurately measuring fetal blood oxygen saturation due to the inherent noise in mixed maternal-fetal photoplethysmogram (PPG) signals, making it difficult to extract the weak fetal signal effectively.
Innovation Solution
A deep neural network-based approach processes mixed PPG signals without explicit fetal signal extraction, utilizing multiple light sources and detectors to modulate distinct wavelengths, and employs a signal quality evaluation module to adaptively adjust sensing hardware, followed by a deep neural network to directly estimate fetal blood oxygen saturation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional signal processing techniques are used to extract fetal signals from mixed maternal-fetal PPG signals, then fetal blood oxygen saturation can be measured, but the measurement accuracy deteriorates due to the weak fetal signal and inherent noise in the mixed signals
Solution Approach 1:
The patent replaces conventional mechanical signal processing techniques with a deep learning-based neural network system. The neural network automatically learns and extracts fetal PPG signals from mixed maternal-fetal signals, eliminating the need for manual signal separation algorithms and improving both accuracy and reliability of fetal SpO2 measurement.
Solution Approach 2:
The patent transforms the approach by changing from direct signal extraction in the time domain to feature representation learning in a higher-dimensional space. The neural network learns optimal parameter transformations and feature extractions automatically, adapting to varying signal conditions and improving measurement robustness.
2Measurement precision
If explicit fetal signal extraction is performed from mixed PPG signals, then fetal SpO2 can be calculated, but the process complexity increases due to the challenging separation of weak fetal signals from noisy maternal signals
Solution Approach 1:
The patent merges signal extraction, noise filtering, and SpO2 calculation into a single integrated deep learning model. The neural network performs multiple functions simultaneously - separating maternal and fetal signals, filtering noise, and estimating SpO2 - thereby reducing overall system complexity while maintaining or improving measurement accuracy.
Solution Approach 2:
The deep learning-based system serves multiple functions: it acts as a signal separator, noise filter, feature extractor, and SpO2 estimator all in one unified framework. This multi-functional approach simplifies the system architecture compared to traditional methods that require separate modules for each function.
3Measurement precision
If multiple light sources with distinct wavelengths are used to improve signal quality, then the accuracy of fetal SpO2 measurement is enhanced, but the device complexity and cost increase
Solution Approach 1:
The patent employs periodic modulation of multiple light sources at distinct wavelengths, allowing the system to encode information from different wavelengths into time-separated signals. This periodic approach enables wavelength multiplexing, improving measurement accuracy without requiring all wavelengths to be active simultaneously, thereby reducing hardware complexity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This method improves the accuracy of fetal blood oxygen saturation estimation by eliminating the need for fetal signal extraction, enhancing signal quality, and providing real-time, accurate fetal SpO2 measurements.
Implementation Method 1
receiving PPG signals from two or more photodetectors detecting transabdominal diffused light from two or more light sources emitting two or more distinct wavelengths
Implementation Method 2
Transabdominal fetal pulse oximetry (TFO) can allow physicians to reliably detect fetal hypoxic distress intrapartum... Measurement of blood oxygen saturation in single-body patients is presently possible through conventional pulse oximetry
Data Source
AI summary
This disclosure provides a fetal-blood-oxygen-saturation estimation technique using a deep neural network without performing explicit fetal signal extraction from mixed maternal-fetal photoplethysmogram (PPG) signals. In one aspect, the disclosed fetal-blood-oxygen-saturation estimation technique receives multiple channels of PPG signals from two or more photodetectors detecting transabdominal diffused light from two or more light sources emitting two or more distinct wavelengths, wherein the photodetectors and light sources are positioned on a maternal abdomen. Note that each channel of the multiple channels of PPG signals includes mixed maternal-fetal PPG signals. Next, the disclosed estimation technique processes the received PPG signals using a trained deep neural network to directly estimate fetal-blood-oxygen-saturation by: tagging the PPG signals with a set of signal-quality levels; feeding the tagged PPG signals as inputs to the deep neural network; and directly inferring, by the deep neural network, fetal-blood-oxygen-saturation estimations and associated confidence levels based on the tagged PPG signals.