Bio-information Estimation Using Differential Signal Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating bio-information, such as blood pressure, from bio-signals face challenges in accurately extracting characteristic points due to noise and instability in pulse waveform signals, particularly in non-ideal environments like motion or light noise.
Innovation Solution
An apparatus and method that utilize a processor to obtain characteristic points from bio-signals by analyzing differential signals, specifically selecting internally dividing points between local minimum and maximum points, and combining these points to estimate bio-information like blood pressure, vascular age, and arterial stiffness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If characteristic points are extracted from bio-signals using conventional methods, then bio-information can be estimated, but the accuracy deteriorates due to noise and instability in pulse waveform signals
Solution Approach 1:
The pulse waveform signal is segmented into multiple components using differential signal analysis. The processor calculates first and second differential signals to identify distinct waveform components (upward convex and downward convex parts), allowing accurate extraction of characteristic points even when the overall signal is noisy or unstable.
Solution Approach 2:
Differential signals serve as an intermediary to bridge the gap between the raw bio-signal and the characteristic points. By computing first and second differential signals, the system creates intermediate representations that highlight specific waveform features while filtering out noise, enabling more reliable extraction of blood pressure and other bio-information.
2Measurement precision
If conventional feature extraction methods are used on bio-signals, then processing can be performed, but measurement accuracy deteriorates in non-ideal environments like motion or light noise
Solution Approach 1:
The system performs preliminary differential signal processing before extracting characteristic points. By pre-calculating first and second differential signals and identifying waveform components in advance, the system prepares a cleaned and structured representation of the bio-signal that is more resistant to subsequent noise interference during measurement.
Solution Approach 2:
The patent replaces direct mechanical signal acquisition with a multi-stage signal processing approach. Instead of relying on the raw mechanical pulse waveform which is susceptible to noise, the system substitutes mathematical differentiation and component analysis to extract features, effectively filtering out motion and light noise through computational methods.
3Measurement precision
If simple characteristic point extraction is used, then processing speed is maintained, but extraction accuracy deteriorates due to inability to identify specific pulse waveform components
Solution Approach 1:
The pulse waveform is segmented into distinct components (upward convex and downward convex parts) through differential signal analysis. This segmentation allows the system to identify and extract characteristic points from specific waveform portions with high accuracy, rather than treating the signal as a whole, while maintaining a systematic processing approach.
Solution Approach 2:
The system transforms the one-dimensional time-series bio-signal into multiple dimensions by computing first and second differential signals. This dimensional transformation creates additional signal representations that reveal waveform component structures invisible in the original signal, enabling more accurate characteristic point extraction despite increased processing steps.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An apparatus for estimating bio-information is disclosed. The bio-information estimating apparatus includes: a sensor configured to measure a bio-signal; and a processor configured to obtain one or more characteristic points, related to one or more pulse waveform components constituting the bio-signal, based on a differential signal of the bio-signal, and to estimate bio-information based on the obtained one or more characteristic points.