Holo-UNet: hologram-to-hologram neural network restoration for high fidelity low light quantitative phase imaging of live cells.

Intensity shot noise in digital holograms distorts the quality of the phase images after phase retrieval, limiting the usefulness of quantitative phase microscopy (QPM) systems in long term live cell imaging. In this paper, we devise a hologram-to-hologram neural network, Holo-UNet, that restores high quality digital holograms under high shot noise conditions (sub-mW/cm2 intensities) at high acquisition rates (sub-milliseconds). In comparison to current phase recovery methods, Holo-UNet denoises the recorded hologram, and so prevents shot noise from propagating through the phase retrieval step that in turn adversely affects phase and intensity images. Holo-UNet was tested on 2 independent QPM systems without any adjustment to the hardware setting. In both cases, Holo-UNet outperformed existing phase recovery and block-matching techniques by ∼ 1.8 folds in phase fidelity as measured by SSIM. Holo-UNet is immediately applicable to a wide range of other high-speed interferometric phase imaging techniques. The network paves the way towards the expansion of high-speed low light QPM biological imaging with minimal dependence on hardware constraints.

Modified inverted selective plane illumination microscopy for sub-micrometer imaging resolution in polydimethylsiloxane soft lithography devices.

Moldable, transparent polydimethylsiloxane (PDMS) elastomer microdevices enable a broad range of complex studies of three-dimensional cellular networks in their microenvironment in vitro. However, the uneven distribution of refractive index change, external to PDMS devices and internally in the sample chamber, creates a significant optical path difference (OPD) that distorts the light sheet beam and so restricts diffraction limited performance. We experimentally showed that an OPD of 120 µm results in the broadening of the lateral point spread function by over 4-fold. In this paper, we demonstrate steps to adapt a commercial inverted selective plane illumination microscope (iSPIM) and remove the OPD so as to achieve sub-micrometer imaging ranging from 0.6 ± 0.04 µm to 0.91 ± 0.03 µm of a fluorescence biological sample suspended in regular saline (RI ≈1.34) enclosed in 1.2 to 2 mm thick micromolded PDMS microdevices. We have proven that the removal of the OPD from the external PDMS layer by refractive index (RI) matching with a readily accessible, inexpensive sucrose solution is critical to achieve a >3-fold imaging resolution improvement. To monitor the RI matching process, a single-mode fiber (SMF) illuminator was integrated into the iSPIM. To remove the OPD inside the PDMS channel, we used an electrically tunable lens (ETL) that par-focuses the light sheet beam with the detection objective lens and so minimised axial distortions to attain sub-micrometer imaging resolution. We termed this new light sheet imaging protocol as modified inverted selective plane illumination microscopy (m-iSPIM). Using the high spatial-temporal 3D imaging of m-iSPIM, we experimentally captured single platelet (≈2 µm) recruitment to a platelet aggregate (22.5 µm × 22.5 µm × 6 µm) under flow at a 150 µm depth within a microfluidic channel. m-iSPIM paves the way for the application of light sheet imaging to a wide range of 3D biological models in microfluidic devices which recapitulate features of the physiological microenvironment and elucidate subcellular responses.