Deep Learning-Based Point-Scanning Super-Resolution Imaging. Fang, L., Monroe, F., Novak, S. W., Kirk, L., Schiavon, C. R., Yu, S. B., Zhang, T., Wu, M., Kastner, K., Kubota, Y., Zhang, Z., Pekkurnaz, G., Mendenhall, J., Harris, K., Howard, J., & Manor, U. bioRxiv, Cold Spring Harbor Laboratory, 2019. ![Deep Learning-Based Point-Scanning Super-Resolution Imaging [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex 2 downloads Point scanning imaging systems (e.g. scanning electron or laser scanning confocal microscopes) are perhaps the most widely used tools for high resolution cellular and tissue imaging. Like all other imaging modalities, the resolution, speed, sample preservation, and signal-to-noise ratio (SNR) of point scanning systems are difficult to optimize simultaneously. In particular, point scanning systems are uniquely constrained by an inverse relationship between imaging speed and pixel resolution. Here we show these limitations can be mitigated via the use of deep learning-based super-sampling of undersampled images acquired on a point-scanning system, which we termed point-scanning super-resolution (PSSR) imaging. Oversampled, high SNR ground truth images acquired on scanning electron or Airyscan laser scanning confocal microscopes were \textquoteleftcrappified\textquoteright to generate semi-synthetic training data for PSSR models that were then used to restore real-world undersampled images. Remarkably, our EM PSSR model could restore undersampled images acquired with different optics, detectors, samples, or sample preparation methods in other labs. PSSR enabled previously unattainable 2 nm resolution images with our serial block face scanning electron microscope system. For fluorescence, we show that undersampled confocal images combined with a multiframe PSSR model trained on Airyscan timelapses facilitates Airyscan-equivalent spatial resolution and SNR with \ 100x lower laser dose and 16x higher frame rates than corresponding high-resolution acquisitions. In conclusion, PSSR facilitates point-scanning image acquisition with otherwise unattainable resolution, speed, and sensitivity.
@article {Fang740548,
author = {Linjing Fang and Fred Monroe and Sammy Weiser Novak and Lyndsey Kirk and Cara R. Schiavon and Seungyoon B. Yu and Tong Zhang and Melissa Wu and Kyle Kastner and Yoshiyuki Kubota and Zhao Zhang and Gulcin Pekkurnaz and John Mendenhall and Kristen Harris and Jeremy Howard and Uri Manor},
title = {Deep Learning-Based Point-Scanning Super-Resolution Imaging},
elocation-id = {740548},
year = {2019},
doi = {10.1101/740548},
publisher = {Cold Spring Harbor Laboratory},
abstract = {Point scanning imaging systems (e.g. scanning electron or laser scanning confocal microscopes) are perhaps the most widely used tools for high resolution cellular and tissue imaging. Like all other imaging modalities, the resolution, speed, sample preservation, and signal-to-noise ratio (SNR) of point scanning systems are difficult to optimize simultaneously. In particular, point scanning systems are uniquely constrained by an inverse relationship between imaging speed and pixel resolution. Here we show these limitations can be mitigated via the use of deep learning-based super-sampling of undersampled images acquired on a point-scanning system, which we termed point-scanning super-resolution (PSSR) imaging. Oversampled, high SNR ground truth images acquired on scanning electron or Airyscan laser scanning confocal microscopes were {\textquoteleft}crappified{\textquoteright} to generate semi-synthetic training data for PSSR models that were then used to restore real-world undersampled images. Remarkably, our EM PSSR model could restore undersampled images acquired with different optics, detectors, samples, or sample preparation methods in other labs. PSSR enabled previously unattainable 2 nm resolution images with our serial block face scanning electron microscope system. For fluorescence, we show that undersampled confocal images combined with a multiframe PSSR model trained on Airyscan timelapses facilitates Airyscan-equivalent spatial resolution and SNR with \~{}100x lower laser dose and 16x higher frame rates than corresponding high-resolution acquisitions. In conclusion, PSSR facilitates point-scanning image acquisition with otherwise unattainable resolution, speed, and sensitivity.},
URL = {https://www.biorxiv.org/content/early/2019/10/23/740548},
eprint = {https://www.biorxiv.org/content/early/2019/10/23/740548.full.pdf},
journal = {bioRxiv}
}
Downloads: 2
{"_id":"r3NkPp3ygN3My7JqR","bibbaseid":"fang-monroe-novak-kirk-schiavon-yu-zhang-wu-etal-deeplearningbasedpointscanningsuperresolutionimaging-2019","author_short":["Fang, L.","Monroe, F.","Novak, S. W.","Kirk, L.","Schiavon, C. R.","Yu, S. B.","Zhang, T.","Wu, M.","Kastner, K.","Kubota, Y.","Zhang, Z.","Pekkurnaz, G.","Mendenhall, J.","Harris, K.","Howard, J.","Manor, U."],"bibdata":{"bibtype":"article","type":"article","author":[{"firstnames":["Linjing"],"propositions":[],"lastnames":["Fang"],"suffixes":[]},{"firstnames":["Fred"],"propositions":[],"lastnames":["Monroe"],"suffixes":[]},{"firstnames":["Sammy","Weiser"],"propositions":[],"lastnames":["Novak"],"suffixes":[]},{"firstnames":["Lyndsey"],"propositions":[],"lastnames":["Kirk"],"suffixes":[]},{"firstnames":["Cara","R."],"propositions":[],"lastnames":["Schiavon"],"suffixes":[]},{"firstnames":["Seungyoon","B."],"propositions":[],"lastnames":["Yu"],"suffixes":[]},{"firstnames":["Tong"],"propositions":[],"lastnames":["Zhang"],"suffixes":[]},{"firstnames":["Melissa"],"propositions":[],"lastnames":["Wu"],"suffixes":[]},{"firstnames":["Kyle"],"propositions":[],"lastnames":["Kastner"],"suffixes":[]},{"firstnames":["Yoshiyuki"],"propositions":[],"lastnames":["Kubota"],"suffixes":[]},{"firstnames":["Zhao"],"propositions":[],"lastnames":["Zhang"],"suffixes":[]},{"firstnames":["Gulcin"],"propositions":[],"lastnames":["Pekkurnaz"],"suffixes":[]},{"firstnames":["John"],"propositions":[],"lastnames":["Mendenhall"],"suffixes":[]},{"firstnames":["Kristen"],"propositions":[],"lastnames":["Harris"],"suffixes":[]},{"firstnames":["Jeremy"],"propositions":[],"lastnames":["Howard"],"suffixes":[]},{"firstnames":["Uri"],"propositions":[],"lastnames":["Manor"],"suffixes":[]}],"title":"Deep Learning-Based Point-Scanning Super-Resolution Imaging","elocation-id":"740548","year":"2019","doi":"10.1101/740548","publisher":"Cold Spring Harbor Laboratory","abstract":"Point scanning imaging systems (e.g. scanning electron or laser scanning confocal microscopes) are perhaps the most widely used tools for high resolution cellular and tissue imaging. Like all other imaging modalities, the resolution, speed, sample preservation, and signal-to-noise ratio (SNR) of point scanning systems are difficult to optimize simultaneously. In particular, point scanning systems are uniquely constrained by an inverse relationship between imaging speed and pixel resolution. Here we show these limitations can be mitigated via the use of deep learning-based super-sampling of undersampled images acquired on a point-scanning system, which we termed point-scanning super-resolution (PSSR) imaging. Oversampled, high SNR ground truth images acquired on scanning electron or Airyscan laser scanning confocal microscopes were \\textquoteleftcrappified\\textquoteright to generate semi-synthetic training data for PSSR models that were then used to restore real-world undersampled images. Remarkably, our EM PSSR model could restore undersampled images acquired with different optics, detectors, samples, or sample preparation methods in other labs. PSSR enabled previously unattainable 2 nm resolution images with our serial block face scanning electron microscope system. For fluorescence, we show that undersampled confocal images combined with a multiframe PSSR model trained on Airyscan timelapses facilitates Airyscan-equivalent spatial resolution and SNR with \\ 100x lower laser dose and 16x higher frame rates than corresponding high-resolution acquisitions. In conclusion, PSSR facilitates point-scanning image acquisition with otherwise unattainable resolution, speed, and sensitivity.","url":"https://www.biorxiv.org/content/early/2019/10/23/740548","eprint":"https://www.biorxiv.org/content/early/2019/10/23/740548.full.pdf","journal":"bioRxiv","bibtex":"@article {Fang740548,\r\n\tauthor = {Linjing Fang and Fred Monroe and Sammy Weiser Novak and Lyndsey Kirk and Cara R. Schiavon and Seungyoon B. Yu and Tong Zhang and Melissa Wu and Kyle Kastner and Yoshiyuki Kubota and Zhao Zhang and Gulcin Pekkurnaz and John Mendenhall and Kristen Harris and Jeremy Howard and Uri Manor},\r\n\ttitle = {Deep Learning-Based Point-Scanning Super-Resolution Imaging},\r\n\telocation-id = {740548},\r\n\tyear = {2019},\r\n\tdoi = {10.1101/740548},\r\n\tpublisher = {Cold Spring Harbor Laboratory},\r\n\tabstract = {Point scanning imaging systems (e.g. scanning electron or laser scanning confocal microscopes) are perhaps the most widely used tools for high resolution cellular and tissue imaging. Like all other imaging modalities, the resolution, speed, sample preservation, and signal-to-noise ratio (SNR) of point scanning systems are difficult to optimize simultaneously. In particular, point scanning systems are uniquely constrained by an inverse relationship between imaging speed and pixel resolution. Here we show these limitations can be mitigated via the use of deep learning-based super-sampling of undersampled images acquired on a point-scanning system, which we termed point-scanning super-resolution (PSSR) imaging. Oversampled, high SNR ground truth images acquired on scanning electron or Airyscan laser scanning confocal microscopes were {\\textquoteleft}crappified{\\textquoteright} to generate semi-synthetic training data for PSSR models that were then used to restore real-world undersampled images. Remarkably, our EM PSSR model could restore undersampled images acquired with different optics, detectors, samples, or sample preparation methods in other labs. PSSR enabled previously unattainable 2 nm resolution images with our serial block face scanning electron microscope system. For fluorescence, we show that undersampled confocal images combined with a multiframe PSSR model trained on Airyscan timelapses facilitates Airyscan-equivalent spatial resolution and SNR with \\~{}100x lower laser dose and 16x higher frame rates than corresponding high-resolution acquisitions. In conclusion, PSSR facilitates point-scanning image acquisition with otherwise unattainable resolution, speed, and sensitivity.},\r\n\tURL = {https://www.biorxiv.org/content/early/2019/10/23/740548},\r\n\teprint = {https://www.biorxiv.org/content/early/2019/10/23/740548.full.pdf},\r\n\tjournal = {bioRxiv}\r\n}\r\n\r\n","author_short":["Fang, L.","Monroe, F.","Novak, S. W.","Kirk, L.","Schiavon, C. R.","Yu, S. B.","Zhang, T.","Wu, M.","Kastner, K.","Kubota, Y.","Zhang, Z.","Pekkurnaz, G.","Mendenhall, J.","Harris, K.","Howard, J.","Manor, U."],"key":"Fang740548","id":"Fang740548","bibbaseid":"fang-monroe-novak-kirk-schiavon-yu-zhang-wu-etal-deeplearningbasedpointscanningsuperresolutionimaging-2019","role":"author","urls":{"Paper":"https://www.biorxiv.org/content/early/2019/10/23/740548"},"metadata":{"authorlinks":{}},"downloads":2},"bibtype":"article","biburl":"https://github.com/salkmanorlab/manor_publications/raw/main/manorlab_pubs.bib","dataSources":["eQQ8KtthHF7NtjjMy","44c9XKrBYFpfipZpH","PGKTciN7m8Gxnak7L","edXsq84pNeYoA6wo6","dr3nCm3wyZkR7nFZf"],"keywords":[],"search_terms":["deep","learning","based","point","scanning","super","resolution","imaging","fang","monroe","novak","kirk","schiavon","yu","zhang","wu","kastner","kubota","zhang","pekkurnaz","mendenhall","harris","howard","manor"],"title":"Deep Learning-Based Point-Scanning Super-Resolution Imaging","year":2019,"downloads":2}