Medical image segmentation method based on mamba network and imaging method
By constructing a medical image segmentation method based on Mamba networks and combining residual connections, convolution, and interactive attention schemes, the method addresses the issues of insufficient reliability and accuracy in existing medical image segmentation technologies. It achieves efficient modeling of complex structures and boundary details, thereby improving segmentation performance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CENT SOUTH UNIV
- Filing Date
- 2025-06-11
- Publication Date
- 2026-07-21
AI Technical Summary
Existing medical image segmentation methods suffer from insufficient reliability and accuracy in medical image segmentation tasks. Furthermore, the existing Vision Mamba architecture is ill-suited to the complex and varied structural morphology and directional heterogeneity of medical images, lacks an effective feature preservation mechanism in deep state modeling, and struggles to achieve semantic coordination between spatially discontinuous regions.
A medical image segmentation method based on Mamba networks is adopted, which combines residual connection scheme, convolution scheme and interactive attention scheme to construct encoding module, graph interactive attention module and decoding module. Through multi-path and multi-directional scanning strategy and channel attention control mechanism, the stability and semantic consistency of feature representation are enhanced, and efficient modeling of complex structures and boundary details is achieved.
It improves the reliability and accuracy of medical image segmentation, enhances the model's ability to represent complex structures and boundary details, and improves segmentation results.
Smart Images

Figure CN120689296B_ABST