Learning the detector in optical tomography

Abstract

We propose a method to reconstruct the optical absorption of a highly-scattering medium probed by diffuse light. The method consists of learning the optical detection system and then using this result to reconstruct the absorption. Our results are illustrated by numerical simulations.

Figures (4)

• Figure 1: Illustrating the experimental geometry.
• Figure 2: Comparison of in-distribution ellipsoid (top two rows) and out-of-distribution breast (bottom two rows) absorptions, with each column showing a cross-section of the three-dimensional absorption image at the indicated height.
• Figure 3: Reconstruction examples for samples from $\mathcal{D}_{\text{ellipsoid}}$, with rows showing the ground truth, Born inversion (BI), CNN, and end-to-end CNN (CNNe2e) results, and columns corresponding to slices at the indicated heights.
• Figure 4: Same as Fig. \ref{['fig:in_distribution3d']}, but with breast data.