FMD-TransUNet: Abdominal Multi-Organ Segmentation Based on Frequency Domain Multi-Axis Representation Learning and Dual Attention Mechanisms
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908547918331904 |
|---|---|
| author | Lu, Fang Xu, Jingyu Sun, Qinxiu Lou, Qiong |
| author_facet | Lu, Fang Xu, Jingyu Sun, Qinxiu Lou, Qiong |
| contents | Accurate abdominal multi-organ segmentation is critical for clinical applications. Although numerous deep learning-based automatic segmentation methods have been developed, they still struggle to segment small, irregular, or anatomically complex organs. Moreover, most current methods focus on spatial-domain analysis, often overlooking the synergistic potential of frequency-domain representations. To address these limitations, we propose a novel framework named FMD-TransUNet for precise abdominal multi-organ segmentation. It innovatively integrates the Multi-axis External Weight Block (MEWB) and the improved dual attention module (DA+) into the TransUNet framework. The MEWB extracts multi-axis frequency-domain features to capture both global anatomical structures and local boundary details, providing complementary information to spatial-domain representations. The DA+ block utilizes depthwise separable convolutions and incorporates spatial and channel attention mechanisms to enhance feature fusion, reduce redundant information, and narrow the semantic gap between the encoder and decoder. Experimental validation on the Synapse dataset shows that FMD-TransUNet outperforms other recent state-of-the-art methods, achieving an average DSC of 81.32\% and a HD of 16.35 mm across eight abdominal organs. Compared to the baseline model, the average DSC increased by 3.84\%, and the average HD decreased by 15.34 mm. These results demonstrate the effectiveness of FMD-TransUNet in improving the accuracy of abdominal multi-organ segmentation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_16044 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FMD-TransUNet: Abdominal Multi-Organ Segmentation Based on Frequency Domain Multi-Axis Representation Learning and Dual Attention Mechanisms Lu, Fang Xu, Jingyu Sun, Qinxiu Lou, Qiong Image and Video Processing Computer Vision and Pattern Recognition Accurate abdominal multi-organ segmentation is critical for clinical applications. Although numerous deep learning-based automatic segmentation methods have been developed, they still struggle to segment small, irregular, or anatomically complex organs. Moreover, most current methods focus on spatial-domain analysis, often overlooking the synergistic potential of frequency-domain representations. To address these limitations, we propose a novel framework named FMD-TransUNet for precise abdominal multi-organ segmentation. It innovatively integrates the Multi-axis External Weight Block (MEWB) and the improved dual attention module (DA+) into the TransUNet framework. The MEWB extracts multi-axis frequency-domain features to capture both global anatomical structures and local boundary details, providing complementary information to spatial-domain representations. The DA+ block utilizes depthwise separable convolutions and incorporates spatial and channel attention mechanisms to enhance feature fusion, reduce redundant information, and narrow the semantic gap between the encoder and decoder. Experimental validation on the Synapse dataset shows that FMD-TransUNet outperforms other recent state-of-the-art methods, achieving an average DSC of 81.32\% and a HD of 16.35 mm across eight abdominal organs. Compared to the baseline model, the average DSC increased by 3.84\%, and the average HD decreased by 15.34 mm. These results demonstrate the effectiveness of FMD-TransUNet in improving the accuracy of abdominal multi-organ segmentation. |
| title | FMD-TransUNet: Abdominal Multi-Organ Segmentation Based on Frequency Domain Multi-Axis Representation Learning and Dual Attention Mechanisms |
| topic | Image and Video Processing Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2509.16044 |