Research Spectral Indices Explained

library(hyperspectR)
#> hyperspectR v0.2.0.9000 - Hyperspectral Imaging Analysis for Biomedical Applications

Hemoglobin Absorption

Tissue oxygenation assessment via HSI relies on the distinct absorption spectra of oxyhemoglobin (HbO2) and deoxyhemoglobin (Hb).

hb_data <- hs_chromophore_data(c("HbO2", "Hb"), wavelength_range = c(430, 910))
library(ggplot2)
ggplot(hb_data, aes(x = wavelength)) +
  geom_line(aes(y = HbO2, color = "HbO2"), linewidth = 0.8) +
  geom_line(aes(y = Hb, color = "Hb"), linewidth = 0.8) +
  scale_color_manual(values = c(HbO2 = "#E41A1C", Hb = "#377EB8")) +
  labs(x = "Wavelength (nm)", y = "Extinction Coefficient",
       title = "Hemoglobin Absorption Spectra", color = "") +
  theme_hsi()

Tissue Oxygen Saturation (StO2)

hs_sto2() reports the fitted HbO2 fraction using tabulated decadic extinction coefficients. It assumes the stated Beer-Lambert model; it is not an empirically validated measure of tissue oxygen saturation for a camera or clinical task.

cube <- hs_example_cube()
sto2 <- hs_sto2(cube)
hs_plot_index(sto2, title = "Fitted Hb fraction (%)", palette = "sto2")

Near-Infrared Perfusion Index (NPI)

The legacy NPI name returns an unscaled NIR band ratio. It has no calibrated perfusion or depth interpretation.

npi <- hs_npi(cube)
hs_plot_index(npi, title = "NIR ratio", palette = "perfusion")

Tissue Hemoglobin Index (THI)

The legacy THI name returns a dimensionless reference/Hb-band ratio. It is not a concentration.

thi <- hs_thi(cube)
hs_plot_index(thi, title = "Hb-band ratio", palette = "hemoglobin")

Custom Normalized Difference Index

Create any two-band ratio index:

ndi <- hs_ndi(cube, band1 = 540, band2 = 660)
hs_plot_index(ndi, title = "NDI (540/660)", range = c(-1, 1))

Research Panel

The research panel combines RGB with model fractions and relative indices. It does not reproduce vendor indices:

hs_plot_clinical(cube)