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

Search research articles

Related Experiment Video

Updated: Jul 4, 2026

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

MAFR-UNet: multi-scale adaptive feature reassembly network for aortic CTA segmentation.

Jian Liu¹, Heqing Huang¹, Yungang Zhang²

¹School of Information Science and Technology, Yunnan Normal University, Kunming, 650500, Yunnan Province, China.

Scientific Reports

|July 2, 2026

Summary

This summary is machine-generated.

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

Clinical Application of Blood Lactate, C-Reactive Protein, and the Age-Adjusted Charlson Comorbidity Index in Predictive Assessment of Malignant Risk of MRSA Pneumonia.

Mediators of inflammation·2026

Same author

A secondary head-to-head comparison of low-intensity focused ultrasound and repetitive transcranial magnetic stimulation for motor recovery after stroke.

PloS one·2026

Same author

Lumican stimulates osteogenic differentiation in bone marrow-derived mesenchymal stem cells through glycolytic reprogramming to alleviate osteoporosis.

Connective tissue research·2025

Same author

Periodically poled thin-film lithium niobate ring Mach Zehnder coupling interferometer as an efficient quantum light source.

Optics express·2025

Same author

Deep Analysis of Imaging Characteristics of Spaceborne SAR Systems as Affected by Antennas Using 3D Antenna Pattern.

Sensors (Basel, Switzerland)·2025

Same author

Reynoutrin activates Sirt1/Foxo3a and Nrf2 pathway via targeting Fyn to ameliorate diabetic nephropathy.

Phytomedicine : international journal of phytotherapy and phytopharmacology·2025

MAFR-UNet improves medical image segmentation by combining Convolutional Neural Networks and Transformers. This hybrid model enhances feature alignment and boundary definition for better multi-organ segmentation accuracy.

Area of Science:

Artificial Intelligence in Medical Imaging
Deep Learning for Medical Image Segmentation
Computer Vision in Healthcare

Background:

Hybrid Convolutional Neural Network (CNN) and Transformer models show promise in medical image segmentation.
Existing methods often suffer from semantic ambiguity due to poor multi-scale feature alignment and noise introduced by direct skip connections.
Transformer models' global representation advantage can be undermined by direct feature transfer in U-shaped architectures.

Purpose of the Study:

To introduce MAFR-UNet, a novel hybrid model integrating CNNs and Transformers for improved medical image segmentation.
To address limitations in multi-scale feature alignment and boundary ambiguity in Computed Tomography Angiography (CTA) images.
To enhance both local and global feature extraction capabilities for robust segmentation.

Keywords:

Adaptive feature reassembly Aortic image segmentation Convolutional neural networks Residual learning Transformer

Related Experiment Videos

Last Updated: Jul 4, 2026

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

Main Methods:

Developed MAFR-UNet, a U-shaped segmentation model combining CNN and Transformer architectures.
Incorporated a Multi-scale Adaptive Feature Reassembly (MAFR) module for multi-level feature capture and alignment.
Integrated a residual CNN bottleneck to strengthen local features and reduce computational cost, alongside a boundary-aware loss term for sharper delineation.

Main Results:

MAFR-UNet achieved competitive segmentation performance on medical imaging tasks.
The model demonstrated effective multi-scale semantic alignment and reduced ambiguity during decoding.
Experimental results showed good generalization capabilities across various multi-organ segmentation tasks.

Conclusions:

MAFR-UNet effectively synergizes CNN and Transformer capabilities for superior medical image segmentation.
The proposed MAFR module and boundary-aware loss significantly improve feature alignment and boundary definition.
The model shows strong potential for diverse clinical applications requiring accurate multi-organ segmentation.