Synthetic Regional Spectral Comparison

R-CMD-check pkgdown CRAN status Codecov test coverage CRAN downloads CRAN downloads total License: MIT Lifecycle: experimental 

library(hyperspectR)

Scope

This example compares regions in a synthetic image. The package has not been validated to diagnose compartment syndrome, detect ischemia, or replace clinical assessment. No clinical threshold is defined by these simulated outputs.

Simulating a Demonstration Scene

The region parameters vary synthetic spectral shapes. Their assigned values are simulation inputs, not established tissue oxygenation ground truth.

cube <- hs_simulate_cube(rows = 50, cols = 50, n_regions = 4,
                          sto2_range = c(0.15, 0.9), seed = 123)

Tissue Oxygenation Mapping

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

The fitted fractions may disagree with simulation parameters because the simulator and reference inversion use different models.

Perfusion Assessment

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

Research Panel Overview

hs_plot_clinical(cube, indices = c("sto2", "npi", "thi"))

ROI Comparison: Affected vs. Unaffected

roi_affected <- hs_roi_rect(cube, x_range = c(30, 50), y_range = c(30, 50))
roi_normal <- hs_roi_rect(cube, x_range = c(1, 20), y_range = c(1, 20))

stats_aff <- hs_roi_stats(cube, roi_affected)
stats_norm <- hs_roi_stats(cube, roi_normal)

library(ggplot2)
ggplot() +
  geom_line(data = stats_norm,
            aes(x = wavelength, y = mean, color = "Normal"), linewidth = 0.8) +
  geom_line(data = stats_aff,
            aes(x = wavelength, y = mean, color = "Affected"), linewidth = 0.8) +
  scale_color_manual(values = c(Normal = "#2E86AB", Affected = "#E41A1C")) +
  labs(x = "Wavelength (nm)", y = "Reflectance",
       title = "Spectral Comparison: Normal vs. Affected Tissue", color = "") +
  theme_hsi()

This plot describes within-image spectral variation. Pixel SD is not a confidence interval for a population effect. For repeated measurements, define subjects as independent groups and use hs_group_summary() on prespecified ROI summaries.