Chinese Journal of Lasers, Volume. 49, Issue 24, 2407206(2022)

Deep Learning in Single-Molecule Localization Microscopy

Tingdan Luo and Yiming Li*
Author Affiliations
  • Department of Biomedical Engineering, Southern University of Science and Technology, Shenzhen 518055, Guangdong, China
  • show less
    Figures & Tables(8)
    Neural network model framework. (a) Multilayer perceptron; (b) convolutional neural network (CNN) based feature network; (c) Inception module; (d) residual network (ResNet); (e) encoder-decoder architecture; (f) long short-term memory (LSTM) network; (g) Transformer; (h) generative adversarial network
    Single-molecule localization method DeepSTORM based on deep learning and simulated dense data[9]. (a) Network architecture of DeepSTORM; (b) diffraction-limited low resolution image (left), DeepSTORM reconstruction with ground truth emitter positions (red crosses) (middle), and magnified view of the selected region (right)
    Image reconstruction processes of DI-STORM and conventional SMLM method[85]
    Design of an optimial PSF using neural networks for multicolor imaging[80]. (a) Amount of phase change of each wavelength after passing through SLMs with same voltage; (b) optimization flow of multi-wavelength PSF based on neural network; (c) optimized SLM voltage pattern for color determination by neural network; (d) phase delay for 565 nm, 625 nm, 705 nm and 800 nm light with same voltage; (e) simulated PSFs of different wavelengths
    Workflow of machine learning-based multidimensional SMLM[10]. (a) Color-separating ANN; (b) ANNs for resolving axial position; (c) analysis of unknown samples
    Data simulation based on SMLM simulator software. (a) Three-dimensional simulation of microtubules by SuReSim[97]; (b) TestSTORM simulated axonal patterns (left) and generated TIFF image (right)[98]; (c) SMeagol simulated fluorescently labeled MinE proteins[99]
    • Table 1. Characteristics of common neural network frameworks

      View table

      Table 1. Characteristics of common neural network frameworks

      ModelCharacteristics
      MLPFully connected class of feedforward ANN
      CNNDeep learning model based on shared-weight architecture of convolution kernels or filters.
      InceptionDeep learning model solving non-uniform sparse data calculation problems by adding 1×1 convolutional layer to reduce dimensionality
      ResNetDeep learning model solving vanishing/exploding gradient problems and improving classification/recognition accuracy
      LSTMSpecial kind of recurrent neural network that is capable of learning long term dependencies in data
      TransformerDeep learning model that adopts self-attention mechanism, differentially weighting significance of each part of input data
      GANDeep learning framework to train generative models via adversarial process
    • Table 2. Common data simulation methods and characteristics

      View table

      Table 2. Common data simulation methods and characteristics

      MethodCharactistics
      SuReSimGenerate simulated localizations from ground truth models in two alternative modes: instant 3D visualization; raw SMLM data in form of image stacks
      TestSTORMGenerate image stacks for conventional localization microscopes with several new features such as scalar and vector diffraction based PSF models, drift, structured background, multicolor imaging, polarization sensitive excitation and detection
      SMeagolGenerate highly realistic single-molecule microscopy time-lapse image series, aiming primarily at single-particle tracking applications
    Tools

    Get Citation

    Copy Citation Text

    Tingdan Luo, Yiming Li. Deep Learning in Single-Molecule Localization Microscopy[J]. Chinese Journal of Lasers, 2022, 49(24): 2407206

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Optical Diagnostics and Therapy

    Received: Aug. 8, 2022

    Accepted: Oct. 8, 2022

    Published Online: Dec. 19, 2022

    The Author Email: Yiming Li (liym2019@sustech.edu.cn)

    DOI:10.3788/CJL202249.2407206

    Topics