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

You might also read

Related Articles

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

Sort by
Same author

Predicting dysthyroid optic neuropathy in moderate-to-severe thyroid eye disease: a clinically applicable nomogram.

European thyroid journal·2025
Same author

The Effects of Counter-Ions on Peptide Structure, Activity, and Applications.

Biomolecules·2025
Same author

Antimicrobial Peptides for Skin Wound Healing.

Biomolecules·2025
Same author

Effects of Maternal Iodine Deficiency in Women with Mild Thyroid Dysfunction During Early Pregnancy on Mental and Motor Growth of Their Offspring During Levothyroxine Treatment.

Biological trace element research·2025
Same author

The Gut Microbiota and Its Metabolites and Their Association with the Risk of Autoimmune Thyroid Disease: A Mendelian Randomization Study.

Nutrients·2024
Same author

Association between the triglyceride-glucose index and thyroid disorders: a cross-sectional survey and Mendelian randomization analysis.

Endocrine·2024

Related Experiment Video

Updated: Aug 19, 2025

Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies
04:25

Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies

Published on: December 15, 2023

2.7K

Learning to segment subcortical structures from noisy annotations with a novel uncertainty-reliability aware learning

Xiang Li1, Ying Wei2, Qian Hu1

  • 1College of Information Science and Engineering, Northeastern University, Shenyang, 110819, China.

Computers in Biology and Medicine
|November 28, 2022
PubMed
Summary
This summary is machine-generated.

This study introduces uncertainty-reliability awareness learning (URAL) for robust brain subcortical structure segmentation using noisy annotations. URAL effectively utilizes all training data, improving segmentation accuracy in medical image analysis.

Keywords:
Contrastive regularizationNoisy labelsPrototypical soft label correctionSubcortical structure segmentationUncertainty-reliability estimation

More Related Videos

A Comprehensive Protocol for Manual Segmentation of the Medial Temporal Lobe Structures
12:30

A Comprehensive Protocol for Manual Segmentation of the Medial Temporal Lobe Structures

Published on: July 2, 2014

20.4K
A Standardized Pipeline for Examining Human Cerebellar Grey Matter Morphometry using Structural Magnetic Resonance Imaging
11:50

A Standardized Pipeline for Examining Human Cerebellar Grey Matter Morphometry using Structural Magnetic Resonance Imaging

Published on: February 4, 2022

4.1K

Related Experiment Videos

Last Updated: Aug 19, 2025

Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies
04:25

Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies

Published on: December 15, 2023

2.7K
A Comprehensive Protocol for Manual Segmentation of the Medial Temporal Lobe Structures
12:30

A Comprehensive Protocol for Manual Segmentation of the Medial Temporal Lobe Structures

Published on: July 2, 2014

20.4K
A Standardized Pipeline for Examining Human Cerebellar Grey Matter Morphometry using Structural Magnetic Resonance Imaging
11:50

A Standardized Pipeline for Examining Human Cerebellar Grey Matter Morphometry using Structural Magnetic Resonance Imaging

Published on: February 4, 2022

4.1K

Area of Science:

  • Neuroimaging
  • Artificial Intelligence
  • Medical Image Analysis

Background:

  • Accurate segmentation of subcortical structures is crucial for quantitative brain image analysis.
  • Convolutional Neural Networks (CNNs) excel in medical image segmentation but struggle with noisy annotations due to annotation acquisition challenges.
  • Existing methods often underutilize training data by selecting only reliably annotated samples, impacting model performance.

Purpose of the Study:

  • To propose a novel robust learning method, Uncertainty-Reliability Awareness Learning (URAL), to effectively utilize all training pixels, even with noisy annotations.
  • To address the limitations of current practices in training segmentation networks with imperfect medical image data.
  • To improve the performance of subcortical structure segmentation models in the presence of label noise.

Main Methods:

  • Introduced Uncertainty-Reliability Awareness Learning (URAL) for robust segmentation.
  • Utilized a meta-learning paradigm with a clean validation set to select reliable training pixels from uncertain predictions.
  • Developed Online Prototypical Soft Label Correction (PSLC) for pseudo-label estimation of unreliable pixels.
  • Integrated segmentation loss for reliable pixels and semi-supervised loss for unreliable pixels.
  • Employed category-wise contrastive regularization for compact feature representation learning.

Main Results:

  • URAL achieved superior Dice scores and Mean Hausdorff Distance (MHD) values on two public brain MRI datasets.
  • The method demonstrated state-of-the-art performance across various label noise settings.
  • URAL effectively leveraged all training pixels, outperforming methods that discard noisy data.

Conclusions:

  • URAL offers a robust and effective approach for brain subcortical structure segmentation with noisy annotations.
  • The proposed method significantly enhances segmentation accuracy by making full use of available training data.
  • URAL represents a significant advancement in training deep learning models for medical image analysis under data imperfections.