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

Age Prediction of Hematoma from Hyperspectral Images Using Convolutional Neural Networks.

Journal of imaging·2026
Same author

Impact of COVID-19 lockdown on weight loss after bariatric surgery: a retrospective single-center study with a follow-up of 3 years.

Langenbeck's archives of surgery·2025
Same author

Analysis of intra- and inter-observer variability in 4D liver ultrasound landmark labeling.

Journal of medical imaging (Bellingham, Wash.)·2025
Same author

Increasing Reachability in Robotic Ultrasound Through Base Placement and Tool Design.

The international journal of medical robotics + computer assisted surgery : MRCAS·2024
Same author

Robotic Assisted Ultrasound-Guided Endovascular Stent Implantation in a Vascular Model.

The international journal of medical robotics + computer assisted surgery : MRCAS·2024
Same author

Corrigendum: SonoBox: development of a robotic ultrasound tomograph for the ultrasound diagnosis of paediatric forearm fractures.

Frontiers in robotics and AI·2024
Same journal

sEEGnal: an automated EEG preprocessing pipeline evaluated against expert-driven preprocessing.

Computers in biology and medicine·2026
Same journal

Corrigendum to "Integrating experimental biology, computational methods, and artificial Intelligence in anticancer drug discovery: Bridging the translational Gap" [Comput. Biol. Med. 213 (2026) 111832].

Computers in biology and medicine·2026
Same journal

Organ dose optimization for a point-of-care forearm X-ray photon-counting CT.

Computers in biology and medicine·2026
Same journal

Physics-guided transformation of breathomic feature spaces into disease-specific representations for respiratory disease classification.

Computers in biology and medicine·2026
Same journal

An AI-driven deep learning pipeline for taxonomic classification and biodiversity assessment of deep-sea environmental DNA.

Computers in biology and medicine·2026
Same journal

Rapid personalisation of cardiovascular models using invasively measured right ventricular pressure.

Computers in biology and medicine·2026
See all related articles

Related Experiment Video

Updated: May 31, 2025

Author Spotlight: Integrating Ultrasound Imaging with Biochemical Markers for Thyroid Disease Diagnosis
05:41

Author Spotlight: Integrating Ultrasound Imaging with Biochemical Markers for Thyroid Disease Diagnosis

Published on: February 9, 2024

524

Real-time deformable structure tracking in 3D ultrasound sequences using deformable convolutional layers.

Daniel Wulff1, Floris Ernst2

  • 1University of Lübeck, Ratzeburger Allee 160, Lübeck, 23562, Schleswig-Holstein, Germany; University of Rostock, Universitätsplatz 1, Rostock, 18055, Mecklenburg-Vorpommern, Germany.

Computers in Biology and Medicine
|January 22, 2025
PubMed
Summary
This summary is machine-generated.

This study introduces a new real-time 3D ultrasound tracking method using deformable convolutions for improved accuracy in radiotherapy guidance. The approach significantly reduces tracking errors and enhances speed for soft tissue motion analysis.

Keywords:
AutoencoderRepresentation learningSonographyUltrasound guidance

More Related Videos

Three-Dimensional Ultrasonic Needle Tip Tracking with a Fiber-Optic Ultrasound Receiver
04:33

Three-Dimensional Ultrasonic Needle Tip Tracking with a Fiber-Optic Ultrasound Receiver

Published on: August 21, 2018

10.3K
3D Ultrasound Imaging: Fast and Cost-effective Morphometry of Musculoskeletal Tissue
08:52

3D Ultrasound Imaging: Fast and Cost-effective Morphometry of Musculoskeletal Tissue

Published on: November 27, 2017

23.1K

Related Experiment Videos

Last Updated: May 31, 2025

Author Spotlight: Integrating Ultrasound Imaging with Biochemical Markers for Thyroid Disease Diagnosis
05:41

Author Spotlight: Integrating Ultrasound Imaging with Biochemical Markers for Thyroid Disease Diagnosis

Published on: February 9, 2024

524
Three-Dimensional Ultrasonic Needle Tip Tracking with a Fiber-Optic Ultrasound Receiver
04:33

Three-Dimensional Ultrasonic Needle Tip Tracking with a Fiber-Optic Ultrasound Receiver

Published on: August 21, 2018

10.3K
3D Ultrasound Imaging: Fast and Cost-effective Morphometry of Musculoskeletal Tissue
08:52

3D Ultrasound Imaging: Fast and Cost-effective Morphometry of Musculoskeletal Tissue

Published on: November 27, 2017

23.1K

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Radiotherapy

Background:

  • 3D ultrasound offers real-time, radiation-free soft tissue imaging, making it valuable for radiotherapy guidance.
  • Accurate real-time tracking of soft tissue motion, including deformation from breathing, is crucial but challenging for therapy.
  • Current methods require robust image analysis to ensure precise target localization during treatment.

Purpose of the Study:

  • To develop a novel, real-time capable 3D ultrasound tracking approach to overcome soft tissue deformation complexities.
  • To enhance the accuracy and efficiency of target tracking for radiotherapy guidance using advanced image analysis.
  • To investigate the efficacy of deformable convolution layers within an autoencoder architecture for deformation-invariant representation learning.

Main Methods:

  • A 3D to 2D ultrasound patch reduction strategy was developed for processing 3D ultrasound data.
  • Deformable convolution layers were integrated into a 2D convolutional autoencoder to learn deformation-invariant features.
  • A greedy local search tracking algorithm was implemented and evaluated on in-vivo 3D liver ultrasound sequences.

Main Results:

  • The proposed tracking approach achieved a mean tracking error of 1.58 ± 0.87 mm, a 10.7% improvement over conventional convolutions.
  • The algorithm demonstrated real-time capability with a mean runtime of approximately 4 ms per 3D ultrasound frame.
  • Deformable convolution layers proved beneficial for learning representations of deformable structures, achieving state-of-the-art accuracy at significantly faster speeds.

Conclusions:

  • Deformable convolution layers enhance the learning of meaningful representations from ultrasound patches, crucial for tracking complex soft tissue motion.
  • The developed real-time tracking method offers a significant speed improvement (up to 100x) over existing techniques while maintaining high accuracy.
  • This advancement holds promise for improving the precision and efficiency of image-guided radiotherapy through robust 3D ultrasound analysis.