@article{ReinhardHelmerichBorasetal.2022,
  author    = {Reinhard, Sebastian and Helmerich, Dominic A. and Boras, Dominik and Sauer, Markus and Kollmannsberger, Philip},
  title     = {ReCSAI: recursive compressed sensing artificial intelligence for confocal lifetime localization microscopy},
  series = {BMC Bioinformatics},
  volume    = {23},
  journal   = {BMC Bioinformatics},
  number    = {1},
  doi       = {10.1186/s12859-022-05071-5},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-299768},
  year      = {2022},
  abstract  = {Background Localization-based super-resolution microscopy resolves macromolecular structures down to a few nanometers by computationally reconstructing fluorescent emitter coordinates from diffraction-limited spots. The most commonly used algorithms are based on fitting parametric models of the point spread function (PSF) to a measured photon distribution. These algorithms make assumptions about the symmetry of the PSF and thus, do not work well with irregular, non-linear PSFs that occur for example in confocal lifetime imaging, where a laser is scanned across the sample. An alternative method for reconstructing sparse emitter sets from noisy, diffraction-limited images is compressed sensing, but due to its high computational cost it has not yet been widely adopted. Deep neural network fitters have recently emerged as a new competitive method for localization microscopy. They can learn to fit arbitrary PSFs, but require extensive simulated training data and do not generalize well. A method to efficiently fit the irregular PSFs from confocal lifetime localization microscopy combining the advantages of deep learning and compressed sensing would greatly improve the acquisition speed and throughput of this method. Results Here we introduce ReCSAI, a compressed sensing neural network to reconstruct localizations for confocal dSTORM, together with a simulation tool to generate training data. We implemented and compared different artificial network architectures, aiming to combine the advantages of compressed sensing and deep learning. We found that a U-Net with a recursive structure inspired by iterative compressed sensing showed the best results on realistic simulated datasets with noise, as well as on real experimentally measured confocal lifetime scanning data. Adding a trainable wavelet denoising layer as prior step further improved the reconstruction quality. Conclusions Our deep learning approach can reach a similar reconstruction accuracy for confocal dSTORM as frame binning with traditional fitting without requiring the acquisition of multiple frames. In addition, our work offers generic insights on the reconstruction of sparse measurements from noisy experimental data by combining compressed sensing and deep learning. We provide the trained networks, the code for network training and inference as well as the simulation tool as python code and Jupyter notebooks for easy reproducibility.},
  language  = {en}
}
@techreport{Gross2022,
  author    = {Groß, Lennart},
  title     = {Advices derived from troubleshooting a sensor-based adaptive optics direct stochastic optical reconstruction microscope},
  doi       = {10.25972/OPUS-28995},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-289951},
  pages     = {20},
  year      = {2022},
  abstract  = {One rarely finds practical guidelines for the implementation of complex optical setups. Here, we aim to provide technical details on the decision making of building and revising a custom sensor-based adaptive optics (AO) direct stochastic optical reconstruction microscope (dSTORM) to provide practical assistance in setting up or troubleshooting similar devices. The foundation of this report is an instrument constructed as part of a master's thesis in 2021, which was built for deep tissue imaging. The setup is presented in the following way: (1) An optical and mechanical overview of the system at the beginning of this internship is given. (2) The optical components are described in detail in the order at which the light passes through, highlighting their working principle and implementation in the system. The optical component include (2A) a focus on even sample illumination, (2B) restoring telecentricity when working with commercial microscope bodies, (2C) the AO elements, namely the deformable mirror (DM) and the wavefront sensor, and their integration, and (2D) the separation of wavefront and image capture using fluorescent beads and a dichroic mirror. After addressing the limitations of the existing setup, modification options are derived. The modifications include the implementation of adjustment only light paths to improve system stability and revise the degrees of freedom of the components and changes in lens choices to meet the specifications of the AO components. Last, the capabilities of the modified setup are presented and discussed: (1) First, we enable epifluorescence imaging of bead samples through 180 µm unstained murine hippocampal tissue with wavefront error correction of ~ 90 \%. Point spread function, wavefront shape and Zernike decomposition of bead samples are presented. (2) Second, we move from epifluorescent to dSTORM imaging of tubulin stained primary mouse hippocampal cells, which are imaged through up to 180 µm of unstained murine hippocampal tissue. We show that full width at half maximum (FWHM) of prominent features can be reduced in size by nearly a magnitude from uncorrected epiflourescence images to dSTORM images corrected by the adaptive optics. We present dSTORM localization count and FWHM of prominent features as as a function of imaging depth.},
  subject      = {Einzelmolek{\"u}lmikroskopie},
  language  = {en}
}