Refractive index reconstruction using optical diffraction tomography algorithms based on single-scattering and multiple-scattering models

Optical diffraction tomography is a popular technique used to retrieve the refractive index (RI) distribution of transparent and refractive specimens. Conventional tomography algorithms based on the first-order Born and Rytov approximations are commonly employed, but are limited to optically thin specimens due to the single-scattering assumption. Tomography reconstruction algorithms incorporating multiple-scattering models can overcome this limitation. However, their iterative nature increases computational cost, making them less suitable for applications requiring higher computational efficiency. In this study, the reconstruction accuracy of two intensity-based optical diffraction tomography algorithms was compared: a transport-of-intensity equation (TIE)-based algorithm that employed a single-scattering model and a Fourier beam propagation (FBP)-based algorithm that incorporated a multiple-scattering model. Both algorithms were applied to reconstruct the RI cross-section of a single-mode fibre core. In addition, two FBP-based reconstruction approaches were compared: digital illumination rotation and digital specimen rotation. The results show that the TIE-based algorithm provides good reconstruction accuracy, but requires additional measures to suppress reconstruction artefacts. The FBP-based algorithm yields higher accuracy, and the digital specimen rotation approach produces more accurate results than the digital illumination rotation approach.