Figure 6

This figure demonstrates how problem parameters impact the error in the attenuation coefficient recovery. For each data point a single A-scan was simulated with 2000 depth measurements. After making the initial DR estimate, the deepest 30% of the pixels were discarded to avoid artifacts. The error (L2) estimates were made for each of the 1400 remaining pixels and averaged. Both the simulations and figure creation were done in Matlab 2019a²⁷, https://www.mathworks.com/. (a) This image shows how the DR reconstruction error varies with different mean attenuation coefficients for a variety of \(\zeta\) values. The \(\zeta\) value does not appear to greatly impact DR Reconstruction fidelity. (b) This image shows how the mean posterior estimator error varies with different mean attenuation coefficients for a variety of \(\zeta\) values. This estimate was computed using Eq. (26). Clearly, the incorporation of of this value into the prior impacts our uncertainty in our Bayesian estimate.