Abstract:Diffuse optical tomography (DOT) utilises near-infrared light for imaging spatially distributed optical parameters, typically the absorption and scattering coefficients. The image reconstruction problem of DOT is an ill-posed inverse problem, due to the non-linear light propagation in tissues and limited boundary measurements. The ill-posedness means that the image reconstruction is sensitive to measurement and modelling errors. The Bayesian approach for the inverse problem of DOT offers the possibility of incorporating prior information about the unknowns, rendering the problem less ill-posed. It also allows marginalisation of modelling errors utilising the so-called Bayesian approximation error method. A more recent trend in image reconstruction techniques is the use of deep learning, which has shown promising results in various applications from image processing to tomographic reconstructions. In this work, we study the non-linear DOT inverse problem of estimating the (absolute) absorption and scattering coefficients utilising a 'model-based' learning approach, essentially intertwining learned components with the model equations of DOT. The proposed approach was validated with 2D simulations and 3D experimental data. We demonstrated improved absorption and scattering estimates for targets with a mix of smooth and sharp image features, implying that the proposed approach could learn image features that are difficult to model using standard Gaussian priors. Furthermore, it was shown that the approach can be utilised in compensating for modelling errors due to coarse discretisation enabling computationally efficient solutions. Overall, the approach provided improved computation times compared to a standard Gauss-Newton iteration.

A Self-supervised Learning Reconstruction Algorithm with an Encoder-Decoder Architecture for Diffuse Optical Tomography

Deep Learning Diffuse Optical Tomography

Model-based Graph Convolutional Network for Diffuse Optical Tomography

Diffusion Equation Engine Deep Learning for Diffuse Optical Tomography

A Novelty Convolutional Neural Network Based Direct Reconstruction for MRI Guided Diffuse Optical Tomography

Back-propagation Neural Network-Based Reconstruction Algorithm for Diffuse Optical Tomography

A Modular Deep Learning-based Approach for Diffuse Optical Tomography Reconstruction

Deep learning-enabled high-speed, multi-parameter diffuse optical tomography

Target Depth-Regularized Reconstruction in Diffuse Optical Tomography Using Ultrasound Segmentation As Prior Information.

Enhanced Model Iteration Algorithm with Graph Neural Network for Diffuse Optical Tomography.

Image Reconstruction for Diffuse Optical Tomography Based on Radiative Transfer Equation

An Em-Like Optimization Scheme For Diffuse Optical Tomography

A Model-Based Iterative Learning Approach for Diffuse Optical Tomography

Difference Imaging from Single Measurements in Diffuse Optical Tomography: a Deep Learning Approach

Optical tomographic imaging for breast cancer detection

Diffuse Optical Tomography Based on Multiple Access Coding

A Compact Breast Shape Acquisition System for Improving Diffuse Optical Tomography Image Reconstructions

Algorithmic depth compensation improves transverse resolution and quantification in functional diffuse optical tomography

Deep learning-based method to accurately estimate breast tissue optical properties in the presence of the chest wall

A combined reconstruction-classification method for diffuse optical tomography

Deep orthogonal multi-wavelength fusion for tomogram-free diagnosis in diffuse optical imaging