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

Combined emission ratiometry and motion tracking for optical mapping of contracting hearts: Validation with monophasic action potentials.

The Journal of physiology·2026
Same author

Oxygen‑sensing histone demethylase KDM6A modulates chondrocyte‑to‑osteoblast transdifferentiation by activating the Wnt/β‑catenin pathway.

International journal of molecular medicine·2026
Same author

KBDR: document retrieval based on graph matching with knowledge enhancement.

Scientific reports·2026
Same author

Heterogeneous biological graph convolutional network for drug-target interaction prediction.

PloS one·2026
Same author

Dynamic prognostic nutritional index and 28-day mortality in sepsis: a joint modeling analysis.

Nutrition & metabolism·2026
Same author

Classification of intestinal inflammation driven by gut microbiota metabolites: a new paradigm for precision treatment of cardiovascular diseases.

Frontiers in microbiology·2026

Related Experiment Video

Updated: Aug 20, 2025

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

481

DBMF: Dual Branch Multiscale Feature Fusion Network for polyp segmentation.

Fangjin Liu1, Zhen Hua1, Jinjiang Li1

  • 1Co-Innovation Center of Shandong Colleges and Universities: Future Intelligent Computing, School of Information and Electronic Engineering, Shandong Technology and Business University, Laishan District, Yantai, 264005, China.

Computers in Biology and Medicine
|November 19, 2022
PubMed
Summary

A new Dual Branch Multiscale Feature Fusion Network (DBMF) improves colorectal polyp segmentation by combining CNN and Transformer features. This method enhances polyp identification and boundary refinement for better colorectal cancer diagnosis.

Keywords:
Attention mechanismConvolutional Neural NetworkMulti-scale feature aggregationPolyp segmentationTransformer

More Related Videos

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

605
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

2.9K

Related Experiment Videos

Last Updated: Aug 20, 2025

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

481
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

605
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

2.9K

Area of Science:

  • Medical Imaging
  • Computer Vision
  • Artificial Intelligence

Background:

  • Accurate colorectal polyp segmentation is crucial for diagnosing and treating colorectal cancer.
  • Existing methods combining Convolutional Neural Networks (CNNs) and Transformers have limitations in integrating local and global features.
  • A need exists for improved polyp segmentation techniques to address challenges in complex scenes and tiny polyp detection.

Purpose of the Study:

  • To introduce a novel Dual Branch Multiscale Feature Fusion Network (DBMF) for enhanced polyp segmentation.
  • To leverage parallel CNN and Transformer branches for comprehensive multi-scale local and global feature extraction.
  • To improve the accuracy and reliability of polyp segmentation for clinical applications.

Main Methods:

  • The proposed DBMF network utilizes parallel CNN and Transformer branches to extract multi-scale local and global contextual information, respectively.
  • A Feature Super Decoder (FSD) fuses multi-level features from both branches to enhance scene parsing and tiny polyp detection.
  • A second parallel decoder (SPD), comprising Multi-scale Feature Aggregation (MFA), Parallel Polarized Self-Attention (PSA), and Reverse Attention Fusion (RAF) modules, refines segmentation boundaries.

Main Results:

  • The DBMF network demonstrated superior performance compared to mainstream polyp segmentation networks across five benchmark datasets (CVC-ClinicDB, Kvasir, CVC-300, CVC-ColonDB, and ETIS).
  • Ablation studies confirmed the effectiveness of the DBMF architecture and its components in improving segmentation accuracy.
  • The DBMF achieved state-of-the-art results in polyp segmentation, outperforming existing methods.

Conclusions:

  • The Dual Branch Multiscale Feature Fusion Network (DBMF) offers a significant advancement in automated polyp segmentation.
  • The parallel fusion of local and global features effectively addresses limitations of previous combined approaches.
  • DBMF shows strong potential for improving diagnostic accuracy in colorectal cancer screening programs.