Table 1 Selected publications from the recent years applying or developing deep learning to microscopy
AuthorsYearDetailsReference
Segmentation/classification
 Ning et al.2005Classification and segmentation of tissues in stages of C. elegans development[61]
 Ciresan et al.2012Segmentation of neuronal membranes in EM[60]
 Ciresan et al.2013Mitosis detection in breast cancer[15]
 Long et al.2015Introduction of fully convolutional networks (fCNN) for segmentation tasks[59]
 Ronneberger et al.2015U-net: significant increase in efficiency for image-to-image network[18]
 Kainz et al.2015Colon gland segmentation, network outperforms handcrafted feature detection[62]
 Kraus et al.2016High-throughput classification and segmentation in yeast[52]
  Xu et al.2016Accelerated cell detection in very large images and images with large cell numbers[46]
 Dürr et al.2016Phenotyping of over 40 000 drug-treated single-cell images[54]
 Van Valen et al.2016Efficient cell–cell segmentation in live-cell imaging[58]
 Richmond et al.2017Identification of cells showing damage from phototoxicity[55]
 Kraus et al.2017Identification of yeast strains and mutants and subcellular protein localizations[53]
 Esteva et al.2017Expert-level classification of skin lesions with mobile phone applicable CNN[14]
 Pärnamaa and Parts2017High-throughput classification of protein localization in yeast[32]
 Godinez et al.2017Phenotyping cells after drug treatment, organelle identification[51]
 Eulenberg et al.2017Identification of cell-cycle phases and differences in disease stages[39]
 Kusumoto et al.2018Detection of epithelial cells derived from iPSCs[56]
 Naylor et al.2018Application of U-net for nuclei segmentation in histopathology sections[69]
 Lu et al.2018Unsupervised learning for protein localization prediction, in human and yeast cells[38]
 Falk et al.2018U-net as ImageJ plugin for non-experts, includes pre-trained network[19]
Artificial labelling
 Christiansen et al.2018Label Prediction in fixed and live cells[21]
 Ounkomol et al.20183D label prediction in live-cell, IF and EM images[20]
Image restoration/super resolution
 Nehme et al.2018SMLM images from diffraction-limited input[81]
 Boyd et al.2018Identifies localization of fluorophores from STORM single frames[82]
 Ouyang et al.2018SMLM reconstruction using very small number of frames to predict SR image[23]
 Nelson and Hess2018PALM reconstruction network trained directly on the image to be analysed[83]
 Weigert et al.2018Denoising and resolution enhancement in different organisms and cell types[22]
 Li et al.2018SMLM localization by deep learning and artefact removal by statistical inference[84]
 Krull et al.2018Image denoising using unsupervised learning on purely noisy input images[41]
 Wang et al.2019Conversion of low NA to high NA or diffraction limited to STED resolution images[24]

This table covers the main four themes where AI has provided solutions to some of the major limitations of microscopy in the recent years.