Linxuan Li, Wenjia Wei, Luyao Yang, Wenwen Zhang, Jiashu Dong, Yahua Liu and Wei Zhao

This is the official implementation of the paper CT-Mamba: A Hybrid Convolutional State Space Model for Low-Dose CT Denoising.

Fig. 1. The overall architecture of the proposed CT-Mamba. (b) The structure of the Multi-Scale Coherence Mamba architecture (MSC-Mamba) in CT-Mamba.

Fig. 2. Detailed structure of the dual-branch Deep NPS Loss and its key components. (a) Structure of the Deep NPS Loss. (b) Detailed structure of the u-Feature Net.

Currently, our source code has not been released, but we plan to open-source it after the paper is officially published, allowing the researchers and developers in the community can reproduce our work and further improve the model.

We will provide detailed installation steps and usage instructions in the upcoming open-source release. The release will include:

• Installation dependencies
• Usage examples
• Training and testing procedures

We welcome your suggestions regarding our code. If you have any questions, feel free to contact us via email at: [zy2219105@buaa.edu.cn] or [weiwenjia@buaa.edu.cn]

@misc{li2024ctmambahybridconvolutionalstate,
      title={CT-Mamba: A Hybrid Convolutional State Space Model for Low-Dose CT Denoising}, 
      author={Linxuan Li and Wenjia Wei and Luyao Yang and Wenwen Zhang and Jiashu Dong and Yahua Liu and Wei Zhao},
      year={2024},
      eprint={2411.07930},
      archivePrefix={arXiv},
      primaryClass={eess.IV},
      url={https://arxiv.org/abs/2411.07930}, 
}