CACFormer: A Hybrid CNN-Transformer Architecture Guided by Channel Attention for 3-D Medical Image Segmentation
IEEE Transactions on Medical Imaging · 2026-11 · 45 (11) · 5321-5335
DOI: 10.1109/tmi.2026.3719644 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Medical Imaging, 2026-11. Verfasst von Hongzhong Tang und 5 weiteren Autor:innen. Hauptthemen laut MEDLINE: Imaging, Three-Dimensional, Convolutional Neural Networks. Gefördert von: Joint Fund of Natural Science of Hunan province, Key Projects of Education Department of Hunan Province, High Performance Computing Platform of Xiangtan University.
Abstract (Originalsprache)
Vision Transformer has achieved significant performance improvements in natural image segmentation tasks owing to its superior global modeling capabilities. However, applying vision Transformers to 3D medical image segmentation is challenging because of the quadratic computational complexity of the self-attention mechanism and their limited generalization on small-scale datasets. To address these limitations, we propose a hybrid CNN-Transformer architecture guided by channel attention, referred to as CACFormer, for 3D medical image segmentation. Specifically, we design a simple and effective channel attention module to guide the fusion of local and global features in each channel. This module adaptively assigns weights to each channel based on its semantic contribution to accurate segmentation. Meanwhile, we introduce a novel linear Transformer variant that integrates a linear attention mechanism with tanh activation. This design encourages the model to focus on the target regions and produce robust segmentation outcomes. The effectiveness and competitive generalization of the proposed framework are validated across five benchmark datasets. On AMOS2022, CACFormer achieves an average Dice score of 89.71%, outperforming 3D UX-Net (89.30%) while reducing inference time from 3.77 s to 2.49 s (a 33.95% reduction). On BraTS2021, CACFormer attains an average Dice score of 90.20%, comparable to TransBTS (90.33%), with 28.54% fewer parameters and 15.22% faster inference time (from 0.46 s to 0.39 s), demonstrating a favorable trade-off between performance and efficiency. Moreover, CACFormer demonstrates competitive cross-dataset generalization, achieving an average Dice score of 86.50% on BraTS2021 when trained on BraTS2019, significantly outperforming TransBTS (47.90%).
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Förderung
- Joint Fund of Natural Science of Hunan province
- Key Projects of Education Department of Hunan Province
- High Performance Computing Platform of Xiangtan University
Zitieren
APA
Tang, H., Zhang, X., Li, Y., Li, C., Liu, T., & Huang, X. (2026). CACFormer: A Hybrid CNN-Transformer Architecture Guided by Channel Attention for 3-D Medical Image Segmentation. IEEE Transactions on Medical Imaging, 45(11), 5321-5335. https://doi.org/10.1109/tmi.2026.3719644
Vancouver
Tang H, Zhang X, Li Y, Li C, Liu T, Huang X. CACFormer: A Hybrid CNN-Transformer Architecture Guided by Channel Attention for 3-D Medical Image Segmentation. IEEE Transactions on Medical Imaging. 2026;45(11):5321-5335. doi:10.1109/tmi.2026.3719644
BibTeX
@article{Tang2026,
author = {Tang, Hongzhong and Zhang, Xiaopeng and Li, Yabo and Li, Chen and Liu, Ting and Huang, Xianghong},
title = {CACFormer: A Hybrid CNN-Transformer Architecture Guided by Channel Attention for 3-D Medical Image Segmentation},
journal = {IEEE Transactions on Medical Imaging},
year = {2026},
volume = {45},
number = {11},
pages = {5321--5335},
doi = {10.1109/tmi.2026.3719644}
}
EUVIMED