Dataset Information

Super-compression of large electron microscopy time series by deep compressive sensing learning.

ABSTRACT: The development of ultrafast detectors for electron microscopy (EM) opens a new door to exploring dynamics of nanomaterials; however, it raises grand challenges for big data processing and storage. Here, we combine deep learning and temporal compressive sensing (TCS) to propose a novel EM big data compression strategy. Specifically, TCS is employed to compress sequential EM images into a single compressed measurement; an end-to-end deep learning network is leveraged to reconstruct the original images. Owing to the significantly improved compression efficiency and built-in denoising capability of the deep learning framework over conventional JPEG compression, compressed videos with a compression ratio of up to 30 can be reconstructed with high fidelity. Using this approach, considerable encoding power, memory, and transmission bandwidth can be saved, allowing it to be deployed to existing detectors. We anticipate the proposed technique will have far-reaching applications in edge computing for EM and other imaging techniques.

SUBMITTER: Zheng S

PROVIDER: S-EPMC8276025 | biostudies-literature | 2021 Jul

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Super-compression of large electron microscopy time series by deep compressive sensing learning.

Zheng Siming S Wang Chunyang C Yuan Xin X Xin Huolin L HL

Patterns (New York, N.Y.) 20210624 7

The development of ultrafast detectors for electron microscopy (EM) opens a new door to exploring dynamics of nanomaterials; however, it raises grand challenges for big data processing and storage. Here, we combine deep learning and temporal compressive sensing (TCS) to propose a novel EM big data compression strategy. Specifically, TCS is employed to compress sequential EM images into a single compressed measurement; an end-to-end deep learning network is leveraged to reconstruct the original i ...[more]

PMID: 34286306

Similar Datasets

Project description:During angiogenesis, growing neovessels must effectively navigate through the tissue space as they elongate and subsequently integrate into a microvascular network. While time series microscopy has provided insight into the cell activities within single growing neovessel sprouts, less is known concerning neovascular dynamics within a large angiogenic tissue bed. Here, we developed a time-lapse imaging technique that allowed visualization and quantification of sprouting neovessels as they form and grow away from adult parent microvessels in three dimensions over cubic millimeters of matrix volume during the course of up to 5 days on the microscope. Using a new image acquisition procedure and novel morphometric analysis tools, we quantified the elongation dynamics of growing neovessels and found an episodic growth pattern accompanied by fluctuations in neovessel diameter. Average elongation rate was 5 μm/h for individual vessels, but we also observed considerable dynamic variability in growth character including retraction and complete regression of entire neovessels. We observed neovessel-to-neovessel directed growth over tens to hundreds of microns preceding tip-to-tip inosculation. As we have previously described via static 3D imaging at discrete time points, we identified different collagen fibril structures associated with the growing neovessel tip and stalk, and observed the coordinated alignment of growing neovessels in a deforming matrix. Overall analysis of the entire image volumes demonstrated that although individual neovessels exhibited episodic growth and regression, there was a monotonic increase in parameters associated with the entire vascular bed such as total network length and number of branch points. This new time-lapse imaging approach corroborated morphometric changes in individual neovessels described by us and others, as well as captured dynamic neovessel behaviors unique to days-long angiogenesis within the forming neovascular network.

Dataset Information

Super-compression of large electron microscopy time series by deep compressive sensing learning.

Publications

Super-compression of large electron microscopy time series by deep compressive sensing learning.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets