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

CRISPR/Cas9 screen identifies DCAF4 as a novel protector of hepatocellular carcinoma against brachytherapy via stress granule-dependent NRF2 activation.

Cell death & disease·2026
Same author

Dynamic nutritional trajectories and deterioration risk in esophageal cancer radiotherapy: a prospective study.

Frontiers in nutrition·2026
Same author

BayeTopo: Bayesian-Based Topology-Guided Learning for Vascular Imaging Segmentation.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
Same author

Cpumycins A-C, Aminated Benzoic Dimers from <i>Streptomyces</i> sp. CPU-W6, and Their Potential Cardioprotective Activities.

Journal of natural products·2026
Same author

Informed-Exploration Reinforcement Learning for Automated Virtual Coronary Intervention Planning.

IEEE transactions on medical imaging·2026
Same author

Knock down LncRNA STARD4-AS1 promotes skin wound healing by inhibiting inflammatory responses through targeting miR-199a-3p.

Burns : journal of the International Society for Burn Injuries·2026

Related Experiment Video

Updated: Oct 25, 2025

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

Multi-level semantic adaptation for few-shot segmentation on cardiac image sequences.

Saidi Guo1, Lin Xu2, Cheng Feng3

  • 1School of Biomedical Engineering, Sun Yat-sen University, China.

Medical Image Analysis
|August 11, 2021
PubMed
Summary

This study introduces Multi-Level Semantic Adaptation (MSA) to improve few-shot segmentation for cardiac imaging. MSA effectively addresses biases in spatial-temporal and long-term information, enhancing accuracy in medical image analysis.

Keywords:
Attention mechanismCardiac image sequencesDomain adaptationFew-shot segmentation

More Related Videos

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

584
Retrospective Cardiac Gating with A Prototype Small-Animal X-ray Computed Tomograph
05:32

Retrospective Cardiac Gating with A Prototype Small-Animal X-ray Computed Tomograph

Published on: February 21, 2025

490

Related Experiment Videos

Last Updated: Oct 25, 2025

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
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

584
Retrospective Cardiac Gating with A Prototype Small-Animal X-ray Computed Tomograph
05:32

Retrospective Cardiac Gating with A Prototype Small-Animal X-ray Computed Tomograph

Published on: February 21, 2025

490

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Computer Vision

Background:

  • Manual segmentation of cardiac image sequences is labor-intensive.
  • Few-shot segmentation offers a solution but faces spatial-temporal and long-term information biases due to the time dimension.
  • These biases lead to over-adaptation in cardiac image sequence analysis.

Purpose of the Study:

  • To propose the Multi-Level Semantic Adaptation (MSA) method for few-shot segmentation of cardiac image sequences.
  • To address spatial-temporal distribution bias and long-term information bias in cardiac image segmentation.
  • To improve the accuracy and efficiency of cardiac image segmentation using limited labeled data.

Main Methods:

  • The proposed MSA method explores domain and weight adaptation on multi-level semantic features (sequence, frame, pixel).
  • Dual-level feature adjustment is used for spatial and temporal domain adaptation, aligning frame and sequence features.
  • Hierarchical attention metric is employed for frame and pixel-level weight adaptation, focusing on relevant features.

Main Results:

  • The MSA method demonstrated effectiveness in few-shot segmentation across three cardiac imaging modalities (MR, CT, Echo).
  • Achieved an average Dice score of 0.9243, indicating high segmentation accuracy.
  • Outperformed ten state-of-the-art methods in few-shot segmentation tasks for cardiac images.

Conclusions:

  • Multi-Level Semantic Adaptation (MSA) is a robust approach for few-shot segmentation of cardiac image sequences.
  • The method successfully mitigates biases inherent in time-series medical imaging data.
  • MSA offers a significant advancement for automated cardiac image analysis, improving diagnostic capabilities.