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

A study on the acceptance intention of digital therapeutics based on UTAUT2 and MIR.

Digital health·2026
Same author

An Effective Microcurrent Stimulation Method for Inducing Non-Pharmacological Parasympathetic Nervous System Activity for Pain Relief.

Bioengineering (Basel, Switzerland)·2026
Same author

Establishing Regional Aβ Cutoffs and Exploring Subgroup Prevalence Across Cognitive Stages Using BeauBrain Amylo<sup>®</sup>.

Dementia and neurocognitive disorders·2025
Same author

Quantifying Brain Atrophy Using a CSF-Focused Segmentation Approach.

Dementia and neurocognitive disorders·2025
Same author

Improving Fall Classification Accuracy of Multi-Input Models Using Three-Axis Accelerometer and Heart Rate Variability Data.

Sensors (Basel, Switzerland)·2025
Same author

Data Reconstruction Methods in Multi-Feature Fusion CNN Model for Enhanced Human Activity Recognition.

Sensors (Basel, Switzerland)·2025

Related Experiment Video

Updated: May 31, 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

2.6K

TransRAUNet: A Deep Neural Network with Reverse Attention Module Using HU Windowing Augmentation for Robust Liver

Kyoung Yoon Lim1, Jae Eun Ko2, Yoo Na Hwang1

  • 1Department of Medical Device and Healthcare, Dongguk University, Seoul 04620, Republic of Korea.

Diagnostics (Basel, Switzerland)
|January 25, 2025
PubMed
Summary

This study introduces TransRAUNet, an AI model for automatic liver vessel segmentation in CT scans. It improves accuracy and efficiency for liver cancer surgeries by enhancing edge detection and context extraction.

Keywords:
CT datasetHounsfield unit windowing augmentationconvolution neural networkdeep learningliver vessel segmentationreverse attention moduletransformer

More Related Videos

DUCT: Double Resin Casting followed by Micro-Computed Tomography for 3D Liver Analysis
12:39

DUCT: Double Resin Casting followed by Micro-Computed Tomography for 3D Liver Analysis

Published on: September 28, 2021

3.3K
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
14:08

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

Published on: April 13, 2013

42.5K

Related Experiment Videos

Last Updated: May 31, 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

2.6K
DUCT: Double Resin Casting followed by Micro-Computed Tomography for 3D Liver Analysis
12:39

DUCT: Double Resin Casting followed by Micro-Computed Tomography for 3D Liver Analysis

Published on: September 28, 2021

3.3K
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
14:08

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

Published on: April 13, 2013

42.5K

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Surgical Planning

Background:

  • Liver cancer poses a global health challenge with high mortality rates.
  • Manual segmentation of liver vessels in CT scans is crucial for surgical planning but is time-consuming and labor-intensive.
  • Developing automated methods for liver vessel segmentation is essential to improve surgical efficiency and patient outcomes.

Purpose of the Study:

  • To develop an accurate and robust deep learning-based method for automatic liver vessel segmentation.
  • To address limitations in current methods, particularly regarding robustness to image variations and segmentation of fine vessel structures.
  • To enhance surgical navigation through precise liver vessel and tumor segmentation.

Main Methods:

  • Proposed TransRAUNet architecture, a modification of TransUNet, incorporating a novel Reverse Attention Module (RAM).
  • Implemented a data augmentation strategy using varying Hounsfield Unit (HU) windowing values to improve robustness to image brightness and contrast.
  • The RAM module was integrated into the upsampling phase to reinforce edge information and minimize false negatives, especially for small vessels.

Main Results:

  • The TransRAUNet model achieved a Dice Similarity Coefficient (DSC) of 0.948 and a sensitivity of 0.944 on the 3Dricadb dataset for liver vessel segmentation.
  • The proposed HU windowing augmentation method demonstrated superior robustness and effectiveness compared to no augmentation and general augmentation techniques.
  • Ablation studies confirmed the performance enhancement provided by the RAM module, outperforming the baseline TransUNet and other state-of-the-art methods.

Conclusions:

  • TransRAUNet offers a significant advancement in automated liver vessel segmentation, improving accuracy and efficiency.
  • The novel augmentation strategy and RAM module contribute to robust performance across varying image conditions.
  • Accurate liver vessel and tumor segmentation using TransRAUNet is expected to aid navigation in liver resection surgeries.