Cargando…
Spectral correction of light emitting diodes enables accurate hydration ratio calculation using narrowband diffuse reflectance spectroscopy
SIGNIFICANCE: Light emitting diodes (LEDs) are commonly utilized for tissue spectroscopy due to their small size, low cost, and simplicity. However, LEDs are often approximated as single-wavelength devices despite having relatively broad spectral bandwidths. When paired with photodiodes, the wavelen...
Autores principales: | , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
Society of Photo-Optical Instrumentation Engineers
2023
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10387445/ https://www.ncbi.nlm.nih.gov/pubmed/37529204 http://dx.doi.org/10.1117/1.JBO.28.7.075005 |
Sumario: | SIGNIFICANCE: Light emitting diodes (LEDs) are commonly utilized for tissue spectroscopy due to their small size, low cost, and simplicity. However, LEDs are often approximated as single-wavelength devices despite having relatively broad spectral bandwidths. When paired with photodiodes, the wavelength information of detected light cannot be resolved. This can result in errors during chromophore concentration calculations. These errors are particularly apparent when analyzing water and fat in the 900 to 1000 nm window where the spectral bandwidth of LEDs can encompass much of the analysis region, resulting in intense crosstalk. AIM: We utilize and present a spectral correction (SC) algorithm to correct for the spectral bandwidth of LEDs. We show the efficacy using a narrowband technique of spectrally broad and overlapping LEDs. APPROACH: Narrowband diffuse reflectance spectroscopy (nb-DRS), a technique capable of quantifying the hydration ratio ([Formula: see text]) of turbid media, was utilized. nb-DRS typically requires a broadband light source and spectrometer. We reduce the hardware to just five LEDs and a photodiode detector, relying on SC to compensate for spectral crosstalk. The effectiveness of our SC approach was tested in simulations as well as in an emulsion phantom and limited selection of human tissue. RESULTS: In simulations, we show that calculated [Formula: see text] errors increased with the spectral bandwidth of LEDs but could be corrected using SC. Likewise, in emulsions, we found an average error of 8.7% (maximum error 14%) if SC was not used. By contrast, applying SC reduced the average error to 2.2% (maximum error of 6.4%). We show that despite utilizing multiple, spectrally broad, and overlapping LEDs, SC was still able to restore the performance of our narrowband method, making it comparable to a much larger full broadband system. |
---|