Fig. 4.

(a) Fractional errors (relative percentage changes) of Model II with respect to Model I over $z/l_t$ when the reduced scattering coefficient ${\mu }_s^{\text{'}}$ is $5$, $10$, $20$ and $30\text{c}{\text{m}}^{-1}$. The common light absorption coefficient ${\mu }_a$ is $0.03\text{c}{\text{m}}^{-1}$. (b) Fractional error over $z/l_t$ when the ${\mu }_a$ value is $0.01$, $0.02$, $0.03$, $0.04$ and $0.05\text{c}{\text{m}}^{-1}$. The common reduced scattering coefficient ${\mu }_s^{\text{'}}$ is $10\text{c}{\text{m}}^{-1}$. (c) Mean fractional error $\overline{\epsilon }(z$, ${\mu }_s^{\text{'}})$ of ${\mu }_{eff}$ estimates with respect to ground truth over 2D domains, axial depth $\text{z}$, and reduced scattering coefficient ${\text{μ}}_{\text{s}}^{\text{'}}$. (d) Mean fractional error $\overline{\epsilon }(z$, ${\mu }_s^{\text{'}})$ over 1D axis $z/l_t$ = $z{\mu }_s^{\text{'}}$. Pixel values of (c) are represented by dots on this graph.