Holotomography (HT) is a powerful label-free imaging technique that enables high-resolution, three-dimensional quantitative phase imaging (QPI) of live cells and organoids through the use of refractive index (RI) distributions as intrinsic imaging contrast1-3. Similar to X-ray computed tomography, HT acquires multiple two-dimensional holograms of a sample at various illumination angles, from which a 3D RI distribution of the sample is reconstructed by inversely solving the wave equation.
   By combining label-free and quantitative 3D imaging capabilities of HT with machine learning approaches, there is potential to provide synergistic capabilities in bioimaging and clinical diagnosis. In this presentation, we will discuss the potential benefits and challenges of combining QPI and artificial intelligence (AI) for various aspects of imaging and analysis, including segmentation, classification, and imaging inference3-6. We will also highlight recent advances in this field and provide insights on future research directions. Overall, the combination of QPI and AI holds great promise for advancing biomedical imaging and diagnostics.

References: 

1. Y. Park, C. Depeursinge and G. Popescu, Nature Photonics 12 (10), 578-589 (2018)

2. Y. Baek and Y. Park, Nature Photonics 15 (5), 354-360 (2021)

3. S. Shin and Y. Park, Nature Materials 21, 317-324 (2022)

4. Y. Jo et al., Nature Cell Biology 23, 1329-1337 (2022)

5. J. Park et al., Nature Methods, 20, pages1645–1660 (2023)

6. G. Kim et al., Nature Methods Review Primers, accepted for publication