Information-rich localization microscopy through machine learning

Kim, Tae-hwan; Moon, Seonah; Xu, Ke

doi:10.1101/373878

“…We use in the simulations 30 × 30 detector's grid patches in which a random number between [5,10] red and green points are located in random positions over the 120 × 120 highresolution grid. Each point is assigned with a random signal value between [6000, 12000 photons].…”

Section: The Recombined Green and Red Channelsmentioning

confidence: 99%

“…For the 2-color simulation [5][6][7][8][9][10], emitters of each color were generated with [5-10K] signal photons and an added background [144,676]. For the 4-color simulation, [30-60K] signal photons were used, and FOVs contained 1-2 emitters of each color (4-8 total) with randomly added background in the range [144,676].…”

Section: Simulated Comparisons Between Psfsmentioning

confidence: 99%

See 1 more Smart Citation

Multicolor localization microscopy and point-spread-function engineering by deep learning

Park

¹

,

Shoman

²

,

Ma

³

et al. 2019

View full text Add to dashboard Cite

Deep learning has become an extremely effective tool for image classification and image restoration problems. Here, we apply deep learning to microscopy and demonstrate how neural networks can exploit the chromatic dependence of the point-spread function to classify the colors of single emitters imaged on a grayscale camera. While existing localization microscopy methods for spectral classification require additional optical elements in the emission path, e.g., spectral filters, prisms, or phase masks, our neural net correctly identifies static and mobile emitters with high efficiency using a standard, unmodified singlechannel configuration. Furthermore, we show how deep learning can be used to design new phase-modulating elements that, when implemented into the imaging path, result in further improved color differentiation between species, including simultaneously differentiating four species in a single image.

show abstract

“…This can be used to monitor the responses of single-molecules to their local environment (Moon et al, 2017) by using fluorophores that are sensitive to pH, ion concentration, or polarity. Machine learning or deep-learning approaches for single-molecule analysis will provide more information on dynamics (Matsunaga & Sugita, 2018), spectral data and three-dimensional (3D) locations (T. Kim, Moon, & Xu, 2018).…”

Section: High-throughput and High-content Single-molecule Assaymentioning

confidence: 99%

Recent advancement of light‐based single‐molecule approaches for studying biomolecules

Croop

¹

,

Zhang

²

,

Lim

³

et al. 2019

WIREs Mechanisms of Disease

View full text Add to dashboard Cite

Recent advances in single‐molecule techniques have led to new discoveries in analytical chemistry, biophysics, and medicine. Understanding the structure and behavior of single biomolecules provides a wealth of information compared to studying large ensembles. However, developing single‐molecule techniques is challenging and requires advances in optics, engineering, biology, and chemistry. In this paper, we will review the state of the art in single‐molecule applications with a focus over the last few years of development. The advancements covered will mainly include light‐based in vitro methods, and we will discuss the fundamentals of each with a focus on the platforms themselves. We will also summarize their limitations and current and future applications to the wider biological and chemical fields. This article is categorized under: Laboratory Methods and Technologies > Imaging Laboratory Methods and Technologies > Macromolecular Interactions, Methods Analytical and Computational Methods > Analytical Methods

show abstract

“…the center of the point-spread function, PSF). It has been shown previously, that in addition to the x-y position, other information can be extracted from the standard PSF as well, such as photophysical properties, emitter depth, orientation, and directionality of motion [9][10][11][12][13].…”

Section: Introductionmentioning

confidence: 99%

Multicolor localization microscopy and point-spread-function engineering by deep learning

Hershko¹,

Weiss²,

Michaeli³

et al. 2019

View full text Add to dashboard Cite

Deep learning has become an extremely effective tool for image classification and image restoration problems. Here, we apply deep learning to microscopy and demonstrate how neural networks can exploit the chromatic dependence of the point-spread function to classify the colors of single emitters imaged on a grayscale camera. While existing localization microscopy methods for spectral classification require additional optical elements in the emission path, e.g., spectral filters, prisms, or phase masks, our neural net correctly identifies static and mobile emitters with high efficiency using a standard, unmodified singlechannel configuration. Furthermore, we show how deep learning can be used to design new phase-modulating elements that, when implemented into the imaging path, result in further improved color differentiation between species, including simultaneously differentiating four species in a single image.

show abstract

Information-rich localization microscopy through machine learning

Cited by 6 publications

References 21 publications

Multicolor localization microscopy and point-spread-function engineering by deep learning

Multicolor localization microscopy and point-spread-function engineering by deep learning

Recent advancement of light‐based single‐molecule approaches for studying biomolecules

Multicolor localization microscopy and point-spread-function engineering by deep learning

Contact Info

Product

Resources

About