IEEE TMI: SegMamba-V2 for general 3D medical image segmentation

Our team published SegMamba-V2 in IEEE Transactions on Medical Imaging, a general framework for long-range modeling in 3D medical image segmentation.

The model extends Mamba-based sequential modeling into volumetric medical imaging and is designed to better capture whole-volume dependencies across different datasets and segmentation targets.

Key points:

• Introduced tri-orientated spatial Mamba blocks for long-range 3D feature modeling.
• Evaluated on CRC-2000 and three additional large-scale 3D segmentation benchmarks.
• Demonstrated strong generalization across modalities, organs, and segmentation targets.

Paper: SegMamba-V2: Long-range Sequential Modeling Mamba For General 3D Medical Image Segmentation