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

Prediction of incident atrial fibrillation from retinal fundus images using a multimodal foundation model.

NPJ digital medicine·2026
Same author

Capturing Finer-grained Long-range Dependency for Dense Prediction in Medical Images: An Empirical Investigation of MLPs.

IEEE journal of biomedical and health informatics·2026
Same author

Hybrid-CMLP: Hybrid CNN-MLP Networks for Low-to-standard-dose PET Synthesis.

IEEE journal of biomedical and health informatics·2026
Same author

Dental Odontogenic Lesion CBCT and Histopathology Integrated Dataset for Benchmarking Deep Learning Algorithms.

Scientific data·2026
Same author

A systematic review of generative artificial intelligence techniques for synthetic medical image datasets: Quality, models, public availability and applications.

Computer methods and programs in biomedicine·2026
Same author

PROTAC-Mediated Targeted Degradation of MDM2 Induces Tumor-Suppressive Signaling in Osteosarcoma Cells.

Cells·2026

Related Experiment Video

Updated: Nov 12, 2025

Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities
07:13

Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities

Published on: October 27, 2023

1.5K

Recurrent feature fusion learning for multi-modality pet-ct tumor segmentation.

Lei Bi1, Michael Fulham2, Nan Li3

  • 1School of Computer Science, University of Sydney, NSW, Australia; Australian Research Council Training Centre for Innovative Bioengineering, NSW, Australia.

Computer Methods and Programs in Biomedicine
|March 21, 2021
PubMed
Summary

This study introduces a novel recurrent fusion network (RFN) for enhanced tumor segmentation in [18f]-fluorodeoxyglucose (FDG) positron emission tomography-computed tomography (PET-CT) scans. The RFN improves segmentation accuracy by iteratively refining multi-modality image features.

Keywords:
Fully convolutional networks (fcns)Positron emission tomography–computed tomography (pet-ct)Segmentation

More Related Videos

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

3.1K
Hybrid µCT-FMT imaging and image analysis
13:45

Hybrid µCT-FMT imaging and image analysis

Published on: June 4, 2015

13.4K

Related Experiment Videos

Last Updated: Nov 12, 2025

Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities
07:13

Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities

Published on: October 27, 2023

1.5K
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

3.1K
Hybrid µCT-FMT imaging and image analysis
13:45

Hybrid µCT-FMT imaging and image analysis

Published on: June 4, 2015

13.4K

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Oncology

Background:

  • [18f]-fluorodeoxyglucose (FDG) PET-CT is crucial for cancer staging, combining metabolic and anatomical information.
  • Automatic tumor segmentation using fully convolutional networks (FCNs) is advancing cancer diagnosis.
  • Existing FCN fusion methods (early, late, hyper-fusion) have limitations in effectively integrating multi-modality image features, potentially leading to inaccurate segmentations, especially for heterogeneous tumors.

Purpose of the Study:

  • To develop an advanced FCN-based method for accurate multi-modality tumor segmentation in PET-CT images.
  • To overcome the limitations of current fusion techniques by proposing a novel recurrent fusion approach.
  • To improve the precision and consistency of automatic tumor segmentation in cancer staging.

Main Methods:

  • Proposed a Recurrent Fusion Network (RFN) employing multiple recurrent fusion phases for progressive integration of multi-modality image features.
  • Implemented iterative learning within recurrent phases to refine segmentation results at each stage.
  • Focused feature learning around intermediary segmentation results to minimize inconsistencies and enhance accuracy.

Main Results:

  • Evaluated the RFN on non-small cell lung cancer (NSCLC) PET-CT datasets, demonstrating superior segmentation accuracy compared to existing methods.
  • Benchmarked against early-fusion, late-fusion, hyper-fusion, and state-of-the-art methods using various network backbones (ResNet, DenseNet, 3D-UNet).
  • Confirmed the generalizability of the RFN across different datasets and network architectures.

Conclusions:

  • Iterative multi-modality feature fusion through recurrent phases significantly refines tumor segmentation results.
  • The proposed RFN consistently produces accurate segmentations across diverse network architectures.
  • This approach offers a more robust and reliable solution for automated tumor segmentation in PET-CT imaging.