Image reconstruction algorithms for optical diffraction tomography microscopy: A review

Optical diffraction tomography (ODT) is a label-free 3D quantitative phase imaging technique that utilizes the refractive index (RI) contrast for visualization, offering advantages such as non-invasiveness, non-destructiveness, long-term imaging, and high spatial resolution (< 200 nm). Over the past decade, significant advances in the ODT system hardware and software have enabled comprehensive visualization of living cells' subcellular structures, driving applications in cell biology, biophysics, and immunology. To fully harness the potential of ODT hardware, it is imperative to develop advanced reconstruction algorithms. This review elaborates on the fundamental physical principles and implementation methods of ODT, with a particular emphasis on image reconstruction algorithms. Various reconstruction approaches are systematically analyzed, including the direct inversion method, the filtered backpropagation method, and various more accurate reconstruction approaches based on high-order scattering models. The strengths and limitations of these methods are summarized, along with discussions on recent advancements in the field. Furthermore, the potential contributions of fluorescence microscopy and deep learning to improving ODT imaging are explored. Finally, future directions for the advancement of ODT technology are proposed.