Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Oct 29, 2025

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

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

Published on: November 30, 2022

Autoencoder based self-supervised test-time adaptation for medical image analysis.

Yufan He¹, Aaron Carass¹, Lianrui Zuo²

¹Dept. of Electrical and Computer Engineering, The Johns Hopkins University, Baltimore, MD 21218, USA.

Medical Image Analysis

|July 10, 2021

Summary

This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

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

Sort by

Same authorSame journal

Beyond the LUMIR challenge: The pathway to foundational registration models.

Medical image analysis·2026

Same author

Metabolomic Signatures of Brain Atrophy and Ibudilast Response in Progressive Multiple Sclerosis.

medRxiv : the preprint server for health sciences·2026

Same author

DSHARP: Deep Incompressible Motion Estimation with Sinusoidal-transformed Harmonic Phase for Tagged MRI.

IEEE transactions on medical imaging·2026

Same author

Editorial for the Special Issue on Harmonization Techniques for MRI.

NeuroImage·2026

Same author

Proteomic Age Acceleration in Multiple Sclerosis Precedes Symptom Onset and Associates with Severity.

medRxiv : the preprint server for health sciences·2026

Same author

Accelerometry-derived activity fragmentation as a predictor of brain atrophy and disability progression in multiple sclerosis.

Multiple sclerosis (Houndmills, Basingstoke, England)·2026

Same journal

ContiMorph: An unsupervised learning framework for cardiac motion tracking with time-continuous diffeomorphism.

Medical image analysis·2026

Same journal

MedP-CLIP: Medical CLIP with region-aware prompt integration.

Medical image analysis·2026

Same journal

Multi-organ guided diagnosis of mild cognitive impairment via hierarchical alignment and knowledge distillation.

Medical image analysis·2026

Same journal

SUDA: Simultaneous unsupervised knowledge distillation and adaptation of foundation models for efficient pathological image analysis.

Medical image analysis·2026

Same journal

Annotation-efficient medical image segmentation via cross-latent graphs and vector-quantized memory.

Medical image analysis·2026

See all related articles

This study introduces a novel deep learning model that rapidly adapts to new medical imaging data domains using only a single test subject. This approach effectively addresses domain shift challenges in clinical practice without costly retraining or access to original training data.

Area of Science:

Medical Image Analysis
Deep Learning
Computer Vision

Background:

Deep neural networks (DNNs) excel in medical image analysis but suffer performance degradation due to domain shift.
Domain shift, caused by variations in data acquisition (e.g., imaging parameters, machines), hinders the clinical deployment of DNNs.
Existing solutions like retraining or data harmonization are often impractical due to data unavailability and computational costs.

Purpose of the Study:

To develop a DNN model that can rapidly adapt to new, unseen data domains during inference.
To overcome user-specific domain shifts without requiring access to the source training data or expensive retraining.
To enable efficient deployment of DNNs in diverse clinical settings with varying data characteristics.

Main Methods:

Keywords:

Medical image analysis Self supervised learning Test time adaptation Unsupervised domain adaptation

Related Experiment Videos

Last Updated: Oct 29, 2025

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

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

Published on: November 30, 2022

Propose a novel neural network architecture comprising a task model (T), autoencoders (AEs), and adaptors (As).
Train the task model and autoencoders on the source dataset.
During inference, adaptors are trained on a single test subject to minimize domain shift via autoencoder reconstruction loss, enabling efficient adaptation.

Main Results:

The proposed model demonstrated significant performance improvements in retinal optical coherence tomography (OCT) image segmentation and MRI T1-to-T2 weighted image synthesis.
Adaptation required minimal computational resources, with only 10 iterations on a single test subject.
The method introduced a small storage overhead (approx. 15 MB for autoencoders).

Conclusions:

The developed self-domain-adapted network effectively addresses domain shift in medical imaging.
The model offers a computationally efficient and practical solution for deploying DNNs in real-world clinical environments.
This approach facilitates the use of advanced AI tools across different healthcare facilities with diverse imaging data.