Generative adversarial DacFormer network for MRI brain tumor segmentation

Abstract Current brain tumor segmentation methods often utilize a U-Net architecture based on efficient convolutional neural networks. While effective, these architectures primarily model local dependencies, lacking the ability to capture global interactions like pure Transformer. However, using pur...

Full description

Saved in:
Bibliographic Details
Main Authors: Muqing Zhang, Qiule Sun, Yutong Han, Mingli Zhang, Wei Wang, Jianxin Zhang
Format: Article
Language:English
Published: Nature Portfolio 2025-05-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-02714-4
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850125860182425600
author Muqing Zhang
Qiule Sun
Yutong Han
Mingli Zhang
Wei Wang
Jianxin Zhang
author_facet Muqing Zhang
Qiule Sun
Yutong Han
Mingli Zhang
Wei Wang
Jianxin Zhang
author_sort Muqing Zhang
collection DOAJ
description Abstract Current brain tumor segmentation methods often utilize a U-Net architecture based on efficient convolutional neural networks. While effective, these architectures primarily model local dependencies, lacking the ability to capture global interactions like pure Transformer. However, using pure Transformer directly causes the network to lose local feature information. To address this limitation, we propose the Generative Adversarial Dilated Attention Convolutional Transformer(GDacFormer). GDacFormer enhances interactions between tumor regions while balancing global and local information through the integration of adversarial learning with an improved transformer module. Specifically, GDacFormer leverages a generative adversarial segmentation network to learn richer and more detailed features. It integrates a novel Transformer module, DacFormer, featuring multi-scale dilated attention and a next convolution block. This module, embedded within the generator, aggregates semantic multi-scale information, efficiently reduces the redundancy in the self-attention mechanism, and enhances local feature representations, thus refining the brain tumor segmentation results. GDacFormer achieves Dice values for whole tumor, core tumor, and enhancing tumor segmentation of 90.9%/90.8%/93.7%, 84.6%/85.7%/93.5%, and 77.9%/79.3%/86.3% on BraTS2019-2021 datasets. Extensive evaluations demonstrate the effectiveness and competitiveness of GDacFormer. The code for GDacFormer will be made publicly available at https://github.com/MuqinZ/GDacFormer.
format Article
id doaj-art-e6351290a3f3406dbc3d3698b7db8a24
institution OA Journals
issn 2045-2322
language English
publishDate 2025-05-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-e6351290a3f3406dbc3d3698b7db8a242025-08-20T02:34:02ZengNature PortfolioScientific Reports2045-23222025-05-0115111210.1038/s41598-025-02714-4Generative adversarial DacFormer network for MRI brain tumor segmentationMuqing Zhang0Qiule Sun1Yutong Han2Mingli Zhang3Wei Wang4Jianxin Zhang5School of Computer Science and Engineering, Dalian Minzu UniversitySchool of Computer Science and Technology, Dalian University of TechnologySchool of Computer Science and Engineering, Dalian Minzu UniversityMontreal Neurological InstituteSchool of Computer Science and Engineering, Dalian Minzu UniversitySchool of Computer Science and Engineering, Dalian Minzu UniversityAbstract Current brain tumor segmentation methods often utilize a U-Net architecture based on efficient convolutional neural networks. While effective, these architectures primarily model local dependencies, lacking the ability to capture global interactions like pure Transformer. However, using pure Transformer directly causes the network to lose local feature information. To address this limitation, we propose the Generative Adversarial Dilated Attention Convolutional Transformer(GDacFormer). GDacFormer enhances interactions between tumor regions while balancing global and local information through the integration of adversarial learning with an improved transformer module. Specifically, GDacFormer leverages a generative adversarial segmentation network to learn richer and more detailed features. It integrates a novel Transformer module, DacFormer, featuring multi-scale dilated attention and a next convolution block. This module, embedded within the generator, aggregates semantic multi-scale information, efficiently reduces the redundancy in the self-attention mechanism, and enhances local feature representations, thus refining the brain tumor segmentation results. GDacFormer achieves Dice values for whole tumor, core tumor, and enhancing tumor segmentation of 90.9%/90.8%/93.7%, 84.6%/85.7%/93.5%, and 77.9%/79.3%/86.3% on BraTS2019-2021 datasets. Extensive evaluations demonstrate the effectiveness and competitiveness of GDacFormer. The code for GDacFormer will be made publicly available at https://github.com/MuqinZ/GDacFormer.https://doi.org/10.1038/s41598-025-02714-4Brain tumor segmentationU-NetGenerative adversarialTransformer
spellingShingle Muqing Zhang
Qiule Sun
Yutong Han
Mingli Zhang
Wei Wang
Jianxin Zhang
Generative adversarial DacFormer network for MRI brain tumor segmentation
Scientific Reports
Brain tumor segmentation
U-Net
Generative adversarial
Transformer
title Generative adversarial DacFormer network for MRI brain tumor segmentation
title_full Generative adversarial DacFormer network for MRI brain tumor segmentation
title_fullStr Generative adversarial DacFormer network for MRI brain tumor segmentation
title_full_unstemmed Generative adversarial DacFormer network for MRI brain tumor segmentation
title_short Generative adversarial DacFormer network for MRI brain tumor segmentation
title_sort generative adversarial dacformer network for mri brain tumor segmentation
topic Brain tumor segmentation
U-Net
Generative adversarial
Transformer
url https://doi.org/10.1038/s41598-025-02714-4
work_keys_str_mv AT muqingzhang generativeadversarialdacformernetworkformribraintumorsegmentation
AT qiulesun generativeadversarialdacformernetworkformribraintumorsegmentation
AT yutonghan generativeadversarialdacformernetworkformribraintumorsegmentation
AT minglizhang generativeadversarialdacformernetworkformribraintumorsegmentation
AT weiwang generativeadversarialdacformernetworkformribraintumorsegmentation
AT jianxinzhang generativeadversarialdacformernetworkformribraintumorsegmentation