Validate multi-distance reconstruction and extend to realistic acquisition conditions

Systematically investigate whether multi-distance time-resolved reflectance measurements improve the retrieval of deep-layer scattering, and extend the learning-based bilayer reconstruction framework to finite instrument response functions, experimental noise, detector effects, and uncertainty in superficial-layer thickness so that it can ultimately be applied to in vivo datasets.

Background

The study evaluates optical-property reconstruction from a single time-resolved reflectance measurement at one source–detector separation under idealized simulation conditions. Although the learning-based method outperforms the analytical diffusion-based inversion overall, the reduced scattering coefficient of the deeper layer remains poorly reconstructed, suggesting limited depth sensitivity in the available measurement.

The authors explicitly identify multi-distance measurements as a potential source of complementary depth information and propose testing whether they improve deep-layer scattering retrieval. They also leave unresolved the method’s robustness under realistic experimental conditions, including nonzero instrument response functions, measurement noise, detector-related effects, and uncertainty in the known superficial-layer thickness. Addressing these issues is presented as necessary for eventual application to in vivo data.