.
Significance
: Biomedical optics system design, image formation, and image analysis have primarily been guided by classical physical modeling and signal processing methodologies. Recently, however, deep learning (DL) has become a major paradigm in computational modeling and has demonstrated utility in numerous scientific domains and various forms of data analysis.
Aim
: We aim to comprehensively review the use of DL applied to macroscopic diffuse optical imaging (DOI).
Approach
: First, we provide a layman introduction to DL. Then, the review summarizes current DL work in some of the most active areas of this field, including optical properties retrieval, fluorescence lifetime imaging, and diffuse optical tomography.
Results
: The advantages of using DL for DOI versus conventional inverse solvers cited in the literature reviewed herein are numerous. These include, among others, a decrease in analysis time (often by many orders of magnitude), increased quantitative reconstruction quality, robustness to noise, and the unique capability to learn complex end-to-end relationships.
Conclusions
: The heavily validated capability of DL’s use across a wide range of complex inverse solving methodologies has enormous potential to bring novel DOI modalities, otherwise deemed impractical for clinical translation, to the patient’s bedside.
Mesoscopic fluorescence molecular tomography (MFMT) is a novel imaging technique capable of obtaining 3-D distribution of molecular probes inside biological tissues at depths of a few millimeters with a resolution up to ~100 μm. However, the ill-conditioned nature of the MFMT inverse problem severely deteriorates its reconstruction performances. Furthermore, dense spatial sampling and fine discretization of the imaging volume required for high resolution reconstructions make the sensitivity matrix (Jacobian) highly correlated, which prevents even advanced algorithms from achieving optimal solutions. In this work, we propose two computational methods to respectively increase the incoherence of the sensitivity matrix and improve the convergence rate of the inverse solver. We first apply a compressed sensing (CS) based preconditioner on either the whole sensitivity matrix or sub sensitivity matrices to reduce the coherence between columns of the sensitivity matrix. Then we employed a regularization method based on the weight iterative improvement method (WIIM) to mitigate the ill-condition of the sensitivity matrix and to drive the iterative optimization process towards convergence at a faster rate. We performed numerical simulations and phantom experiments to validate the effectiveness of the proposed strategies. In both and cases, we were able to improve the quality of MFMT reconstructions significantly.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.