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

Arsenic trioxide depletes cancer stem-like cells and inhibits repopulation of neurosphere derived from glioblastoma by downregulation of Notch pathway.

Toxicology letters·2013
Same author

A prospective, randomized, open-label study comparing the efficacy and safety of preprandial and prandial insulin in combination with acarbose in elderly, insulin-requiring patients with type 2 diabetes mellitus.

Diabetes technology & therapeutics·2013
Same author

Synthesis of the C-18-C-34 fragment of amphidinolides C, C2, and C3.

Organic letters·2013
Same author

Synthesis of the C-1-C-17 fragment of amphidinolides C, C2, C3, and F.

Organic letters·2013
Same author

77Se solid-state NMR of As2Se3, As4Se4 and As4Se3 crystals: a combined experimental and computational study.

Physical chemistry chemical physics : PCCP·2013
Same author

Nanocellulose electroconductive composites.

Nanoscale·2013

Related Experiment Video

Updated: Jan 8, 2026

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

3.3K

Test-time generative augmentation for medical image segmentation.

Xiao Ma1, Yuhui Tao2, Zetian Zhang3

  • 1School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, China; Bioengineering Department and Imperial-X, Imperial College London, London, W12 7SL, UK.

Medical Image Analysis
|December 13, 2025
PubMed
Summary
This summary is machine-generated.

Test-Time Generative Augmentation (TTGA) enhances medical image segmentation by creating diverse, context-aware augmentations during inference. This novel approach improves accuracy and provides pixel-wise error estimation for better clinical insights.

Keywords:
Generative modelMedical image segmentationTest-time augmentation

More Related Videos

Author Spotlight: Segmentation and VR for Advanced Neurovascular Interventions
06:18

Author Spotlight: Segmentation and VR for Advanced Neurovascular Interventions

Published on: April 5, 2024

1.5K
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

723

Related Experiment Videos

Last Updated: Jan 8, 2026

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

3.3K
Author Spotlight: Segmentation and VR for Advanced Neurovascular Interventions
06:18

Author Spotlight: Segmentation and VR for Advanced Neurovascular Interventions

Published on: April 5, 2024

1.5K
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

723

Area of Science:

  • Medical Imaging
  • Computer Vision
  • Artificial Intelligence

Background:

  • Medical image segmentation is vital for clinical applications but faces challenges like occlusions and boundary ambiguities.
  • Existing test-time augmentation (TTA) methods are limited by predefined transformations, hindering adaptability in complex medical imaging scenarios.

Purpose of the Study:

  • To introduce Test-Time Generative Augmentation (TTGA), a novel strategy for enhancing medical image segmentation at inference time.
  • To address limitations of conventional augmentation techniques by providing contextually relevant and diverse augmentations tailored to individual test images.

Main Methods:

  • Developed TTGA leveraging a domain-fine-tuned generative model, specifically diffusion model inversion with masked null-text inversion for region-specific augmentations.
  • Implemented a dual denoising pathway to balance image identity preservation with controlled variability during augmentation generation.
  • Validated TTGA across three segmentation tasks and nine datasets.

Main Results:

  • TTGA significantly improved segmentation accuracy, showing Dice Similarity Coefficient (DSC) gains from 0.1% to 2.3% over baseline methods.
  • The method provided pixel-wise error estimation, achieving DSC gains from 1.1% to 29.0% over baseline.
  • Consistent performance improvements were observed across diverse medical imaging segmentation tasks.

Conclusions:

  • TTGA represents a significant advancement in medical image segmentation by enabling adaptive, generative augmentation at inference time.
  • The proposed method enhances segmentation accuracy and offers valuable pixel-wise error estimation, contributing to more reliable clinical decision-making.
  • The study provides an open-source implementation, facilitating further research and application of TTGA in medical image analysis.