Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

Super-resolution Fluorescence Microscopy01:37

Super-resolution Fluorescence Microscopy

8.2K
Super-resolution fluorescence microscopy (SRFM) provides a better resolution than conventional fluorescence microscopy by reducing the point spread function (PSF). PSF is the light intensity distribution from a point that causes it to appear blurred. Due to PSF, each fluorescing point appears bigger than its actual size, and it is the PSF interference of nearby fluorophores that causes the blurred image. Various approaches to achieving higher resolution through SRFM have recently been...
8.2K

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Impact of training with a 3-dimensional printed cystic duct malformation model on laparoscopic cholecystectomy procedures.

Updates in surgery·2026
Same author

Evolocumab protects mice from radiation-induced damage by affecting hematopoietic cell lineage.

Genomics·2026
Same author

Causal and directional elements of global brain dynamics.

bioRxiv : the preprint server for biology·2026
Same author

Characterization of time-dependent x-ray drive at the center of a cylindrical hohlraum.

Physical review. E·2026
Same author

IGLC3<sup>-</sup> tumor cells drive chemoresistance in colorectal cancer by polarizing SPP1<sup>+</sup> macrophages via the CD44-Wnt-BTF3 axis.

Frontiers in immunology·2026
Same author

LTF-MSPCNet: A synergistic approach combining attention mechanisms and local texture features for oil spill segmentation in SAR images.

Marine pollution bulletin·2026

Related Experiment Video

Updated: Sep 25, 2025

Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

9.4K

U-RISC: An Annotated Ultra-High-Resolution Electron Microscopy Dataset Challenging the Existing Deep Learning

Ruohua Shi1,2, Wenyao Wang1, Zhixuan Li2

  • 1Beijing Academy of Artificial Intelligence, Beijing, China.

Frontiers in Computational Neuroscience
|April 28, 2022
PubMed
Summary

We introduce U-RISC, the largest cell membrane-annotated electron microscopy dataset, to advance connectomics. Current deep learning methods show a significant gap compared to human performance in segmenting neural connections.

Keywords:
EM datasetautomatic cell segmentationconnectomicsdeep learningtransfer learning

More Related Videos

High-Speed Ultraviolet Photoacoustic Microscopy for Histological Imaging with Virtual-Staining assisted by Deep Learning
09:31

High-Speed Ultraviolet Photoacoustic Microscopy for Histological Imaging with Virtual-Staining assisted by Deep Learning

Published on: April 28, 2022

3.2K
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.0K

Related Experiment Videos

Last Updated: Sep 25, 2025

Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

9.4K
High-Speed Ultraviolet Photoacoustic Microscopy for Histological Imaging with Virtual-Staining assisted by Deep Learning
09:31

High-Speed Ultraviolet Photoacoustic Microscopy for Histological Imaging with Virtual-Staining assisted by Deep Learning

Published on: April 28, 2022

3.2K
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.0K

Area of Science:

  • Neuroscience and Computational Biology
  • Connectomics and Neural Circuit Reconstruction
  • Advanced Image Analysis and Deep Learning

Background:

  • Connectomics aims to map neural connections at nanometer scale, crucial for understanding brain function.
  • Deep learning significantly advanced neural connectomic data analysis, yet current methods struggle to meet research demands.
  • Existing datasets and benchmarks do not fully capture the complexity of high-resolution electron microscopy data for cell membrane segmentation.

Purpose of the Study:

  • To introduce the U-RISC dataset, the largest annotated electron microscopy dataset for cell membrane segmentation.
  • To evaluate the performance gap of current deep learning methods against human-level performance on ultra-high resolution EM data.
  • To provide insights into the limitations of deep learning for cell membrane segmentation and establish a new benchmark.

Main Methods:

  • Development and curation of the U-RISC dataset: an ultra-high resolution (2.18 nm/pixel) annotated electron microscopy dataset.
  • An open competition to assess state-of-the-art deep learning methods on the U-RISC dataset.
  • Attribution analysis to investigate the reasons behind deep learning model performance discrepancies.

Main Results:

  • U-RISC is the largest cell membrane-annotated EM dataset, featuring iterative annotations for high quality.
  • Deep learning methods showed a considerable performance gap compared to human-level segmentation on U-RISC, unlike on ISBI 2012.
  • Attribution analysis revealed that U-RISC requires larger contextual information for accurate pixel prediction.
  • A new benchmark of 0.67 was established, surpassing the competition leader by 10%.

Conclusions:

  • The U-RISC dataset presents a significant challenge for current deep learning models in cell membrane segmentation.
  • Further development of deep learning algorithms is needed to bridge the performance gap in high-resolution connectomic data analysis.
  • The U-RISC dataset and associated code are publicly available to foster research in connectomics and AI.