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

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

Chikungunya resurgence highlights gaps in <i>Aedes</i> surveillance and control in South America.

Lancet regional health. Americas·2026

Same author

Case-based evidence links increased SARS-CoV-2 diversity to impaired IFN-I response in severe COVID-19.

Journal of human immunity·2026

Same author

Evaluation of a quasi-automatic treatment planning workflow for head-and-neck cancer in radiotherapy.

Physics and imaging in radiation oncology·2026

Same author

Free flight angular acceptance (FFAA) variance reduction technique for SPECT Monte Carlo simulations.

Physics in medicine and biology·2026

Same author

Image pre-processing impact on generative model performance for Unsupervised Venous Malformation Segmentation.

Computer methods and programs in biomedicine·2026

Same author

Patient-specific lung simulation incorporating regional elastance and recruitment to guide mechanical ventilation in acute respiratory distress syndrome.

Journal of applied physiology (Bethesda, Md. : 1985)·2026

Search research articles

Related Experiment Video

Updated: Oct 13, 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

Improving motion-mask segmentation in thoracic CT with multiplanar U-nets.

Ludmilla Penarrubia¹, Nicolas Pinon¹, Emmanuel Roux¹

¹Univ Lyon, Université Claude Bernard Lyon 1, INSA-Lyon, CNRS, Inserm, CREATIS UMR 5220, U1294, F-69621, Lyon, France.

Medical Physics

|November 15, 2021

Summary

This summary is machine-generated.

A new deep learning method accurately segments lung motion masks from CT scans, improving image registration for radiotherapy and ventilation analysis. This robust approach is fast and suitable for clinical use.

Keywords:

deep learning segmentation thoracic CT

More Related Videos

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

Related Experiment Videos

Last Updated: Oct 13, 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

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

Area of Science:

Medical Imaging
Deep Learning
Computational Anatomy

Background:

Motion-mask segmentation in thoracic CT identifies lung and visceral regions with significant breathing-induced displacements.
Accurate motion masks are crucial for inter-phase CT image registration, aiding radiotherapy planning and local lung ventilation calculations.
Current methods lack the robustness for routine clinical application, highlighting a need for improved segmentation techniques.

Purpose of the Study:

To develop and validate a robust, lightweight deep learning approach for motion-mask segmentation in thoracic CT images.
To demonstrate the feasibility of achieving high accuracy without data augmentation or complex model designs on standard computing hardware.

Main Methods:

A convolutional neural network (CNN) architecture employing three 2D U-Nets (sagittal, coronal, axial) was utilized.
Predictions from the U-Nets were combined via majority voting to generate a 3D motion mask.
The model was trained and evaluated using K-fold cross-validation on 4D CT datasets from lung cancer patients, with generalizability assessed on a separate COVID-19 patient dataset.

Main Results:

The deep learning approach significantly outperformed a baseline level-set method, increasing the success rate and eliminating failures.
The method achieved substantial speed-up factors (60x with GPU, 17x with CPU) and maintained a low memory footprint.
Generalizability was confirmed on an unseen dataset, showing improved success rates across diverse CT characteristics.

Conclusions:

The proposed deep learning method offers a fast (5-second processing on mid-range GPU) and robust solution for motion-mask segmentation, suitable for clinical practice.
Further improvements in success rates are achievable through data augmentation and incorporating diverse annotated data.
The developed code and trained model are publicly available to facilitate clinical adoption and further research.