LED-Photodiode Spectral Correction for Physiological Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing narrowband diffuse reflectance spectroscopy (nb-DRS) techniques using LEDs and photodiodes face challenges with spectral cross-talk due to the wide spectral bandwidth of LEDs, leading to errors in calculating hydration ratios (RH2O) and loss of high-resolution spectrum data, particularly in the 900-1000 nm range.
Innovation Solution
A spectral correction method is applied to correct the spectral bandwidth of LEDs, using a diffusion model and normalized weight data to calculate correction data, reducing errors and improving accuracy in estimating physiological information such as water and fat concentrations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If multi-wavelength LEDs are used as light sources instead of broadband lamps, then device miniaturization and cost reduction are achieved, but spectral cross-talk occurs due to the wide FWHM of LEDs
Solution Approach 1:
A correction profile is introduced as an intermediary element that mediates between the LED spectral characteristics and the photodetector measurements. The correction profile, stored in memory, contains wavelength-specific correction factors that compensate for the LED's wide spectral bandwidth effects, allowing accurate physiological information extraction despite using compact LED-based illumination
Solution Approach 2:
The system changes the operational parameters by using multiple LEDs with different peak wavelengths (e.g., 907nm, 940nm, 980nm) and applies corresponding wavelength-specific correction factors from the correction profile. This parameter-based approach allows the system to operate in the narrowband regime effectively, compensating for each LED's spectral characteristics through software-based correction
2Device complexity
If photodiodes are used instead of spectrometers to detect light, then device complexity is reduced, but high-resolution spectrum data is lost
Solution Approach 1:
Instead of using a spectrometer to directly measure the spectrum, the system creates a computational copy of the spectral information by applying correction profile factors to the photodiode measurements. The correction profile essentially copies the spectral characteristics that would have been measured by a spectrometer, allowing the system to reconstruct spectral information from simpler photodiode readings
Solution Approach 2:
The correction profile acts as an intermediary that bridges the gap between the simple photodiode detection and the need for spectral information. By multiplying the photodiode signal by wavelength-specific correction factors from the profile, the system recovers spectral information without requiring a spectrometer
3Measurement precision
If correction profile calculation is performed using diffusion model and simulated data, then measurement accuracy is improved, but computational complexity increases
Solution Approach 1:
The correction profile is calculated in advance using diffusion models and simulated data, storing the computationally intensive work in a pre-computed lookup table. During actual measurements, the system only needs to retrieve and apply the pre-calculated correction factors, significantly reducing the computational burden during real-time operation while maintaining high measurement accuracy
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
The method effectively reduces errors from spectral cross-talk, achieving an average error of 2.2% compared to 8.7% without correction, restoring the performance of narrowband methods to a level similar to full broadband systems.
Implementation Method 1
receiving, by photodiodes, a first optical data set associated with light irradiated by plural light emitting diodes (LEDs) to a first substance
Implementation Method 2
the diffusion model may be a model that predicts a diffuse reflectance for a light-irradiated medium based on a chromophore concentration and an extinction coefficient
Implementation Method 3
receiving, by photodiodes, a first optical data set associated with light irradiated by plural light emitting diodes (LEDs)
Data Source
AI summary
The present disclosure provides a method for estimating physiological information performed by at least one processor. This method may include: receiving, by photodiodes, a first optical data set associated with light irradiated by plural light emitting diodes (LEDs) to a first substance; receiving, by the photodiodes, a second optical data set associated with light irradiated by the plural LEDs to a second substance; calculating first correction data associated with the plural LEDs based on the first optical data set and the second optical data set; calculating second correction data associated with the plural LEDs based on the first correction data and profile data associated with the plural LEDs; and estimating physiological information associated with the second substance based on the first correction data and the second correction data.