Skip to content

Official implementation of "PanopMamba: Vision State Space Modeling for Nuclei Panoptic Segmentation".

License

Notifications You must be signed in to change notification settings

mkang315/PanopMamba

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

83 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Official PanopMamba, image-level Panoptic Quality (iPQ), boundary-weighted Panoptic Quality (wPQ), and frequency-weighted Panoptic Quality (fwPQ)

Build Build

Description

This is the source code for the paper, "PanopMamba: Vision State Space Modeling for Nuclei Panoptic Segmentation", of which I am the first author. The paper is a preprint without the intention to publish and available to download from arXiv.

Model

Our proposed model is built on Python package MMSegmentation v1.2.2 (released on 14 Dec 2023). The model configuration file is in the folder ./configs/panopmamba and dataset configuration file is in ./mmseg/datasets. The pretrained MSVMamba checkpoint is provided in the MSVMamba repo. The modules are composed by models in ./mmseg/models. All the evaluation metrics, including the proposed iPQ, wPQ, and fwPQ, are in ./tools/evaluation.

Installation

Install requirements.txt in a Python >= 3.8 environment, including PyTorch >= 1.12.

pip install -r requirements.txt

Training CLI

python /tools/train.py

Testing CLI

python /tools/test.py

For further information, please see user guides for the basic usage of MMSegmentation. There are also advanced tutorials for in-depth understanding of mmseg design and implementation.

Evaluation

We trained and evaluated PanopMamba on the Hematoxylin and Eosin (H&E) stained datasets MoNuSAC2020 and NuInsSeg. Here are some addtional experimental results.

Referencing Guide

Please cite the paper if using this repository. Here is a guide to referencing this work in various styles for formatting your references:

Plain Text

  • IEEE Reference Style
    M. Kang, F. F. Ting, R. C.-W. Phan, Z. Ge, and C.-M. Ting, "Panopmamba: Vision state space modeling for nuclei panoptic segmentation," 2026, arXiv:2601.16631.

  • IEEE Full Name Reference Style
    M. Kang, F. F. Ting, R. C.-W. Phan, Z. Ge, and C.-M. Ting, "Panopmamba: Vision state space modeling for nuclei panoptic segmentation," arXiv:2601.16631, 2026.
    NOTE: This is a modification to the standard IEEE Reference Style and used by most IEEE/CVF conferences, including CVPR, ICCV, and WACV, to render first names in the bibliography as "Firstname Lastname" rather than "F. Lastname" or "Lastname, F.".
     - IJCAI Full Name-Year Variation
    [Kang et al., 2026] Ming Kang, Fung Fung Ting, Raphaël C.-W. Phan, Zongyuan Ge, and Chee-Ming Ting. Panopmamba: Vision state space modeling for nuclei panoptic segmentation. arXiv preprint arXiv:2601.16631, 2026.
     - ACL Full Name-Year Variation
    Ming Kang, Fung Fung Ting, Raphaël C.-W. Phan, Zongyuan Ge, and Chee-Ming Ting. 2026. Panopmamba: Vision state space modeling for nuclei panoptic segmentation. arXiv preprint arXiv:2601.16631.

  • Nature Reference Style
    Kang, M., Ting, F. F., Phan, R. C.-W., Ge, Z. & Ting, C.-M. PanopMamba: vision state space modeling for nuclei panoptic segmentation. Preprint at https://arxiv.org/abs/2601.16631 (2026).

  • Springer Reference Style
    Kang, M., Ting, F.F., Phan, R.C.-W., Ge, Z., Ting, C.-M.: PanopMamba: vision state space modeling for nuclei panoptic segmentation (2026). arXiv preprint arXiv:2601.16631

  • Elsevier Numbered Style
    M. Kang, F.F. Ting, R.C.-W. Phan, Z. Ge, C.-M. Ting, PanopMamba: vision state space modeling for nuclei panoptic segmentation, arXiv preprint arXiv:2601.16631 (2026).

  • Elsevier Name–Date (Harvard) Style
    Kang, M., Ting, F.F., Phan, R.C.-W., Ge, Z., Ting, C.-M., 2026. PanopMamba: vision state space modeling for nuclei panoptic segmentation. arXiv preprint arXiv:2601.16631.

  • Elsevier Vancouver Style
    Kang M, Ting FF, Phan RC-W, Ge Z, Ting C-M. PanopMamba: vision state space modeling for nuclei panoptic segmentation. ArXiv [Preprint]. 2026 arXiv:2601.16631 [posted 2026 Jan 23]: [10 p.]. Available from: https://arxiv.org/abs/2601.16631 doi: https://doi.org/10.48550/arXiv.2601.16631

  • Elsevier Embellished Vancouver Style
    Kang M, Ting FF, Phan RC-W, Ge Z, Ting C-M. PanopMamba: vision state space modeling for nuclei panoptic segmentation. ArXiv [Preprint]. 2026 arXiv:2601.16631 [posted 2026 Jan 23]: [10 p.]. Available from: https://arxiv.org/abs/2601.16631 doi: https://doi.org/10.48550/arXiv.2601.16631

  • APA7 (Author–Date) Style
    Kang, M., Ting, F. F., Phan, R. C.-W., Ge, Z., & Ting, C.-M. (2026). PanopMamba: Vision state space modeling for nuclei panoptic segmentation. ArXiv, https://doi.org/10.48550/arXiv.2601.16631
     - AAAI (Author–Year) Variation
    Kang, M.; Ting, F. F.; Phan, R. C.-W.; Ge, Z.; and Ting, C.-M. 2026. PanopMamba: Vision State Space Modeling for Nuclei Panoptic Segmentation. arXiv:2601.16631.
     - ICML (Author–Year) Variation
    Kang, M., Ting, F. F., Phan, R. C.-W., Ge, Z., and Ting, C.-M. PanopMamba: Vision state space modeling for nuclei panoptic segmentation. arXiv preprint arXiv:2601.16631, 2026.
    NOTE: For NeurIPS and ICLR, any reference/citation style is acceptable as long as it is used consistently. The sample of references in Formatting Instructions For NeurIPS almost follows APA7 (author–date) style and that in Formatting Instructions For ICLR Conference Submissions is similar to IJCAI full name-year variation.

BibTeX Format

\begin{thebibliography}{1}
\bibitem{bib1} M. Kang, R. C.-W. Phan, F. F. Ting, Z. Ge, and C.-M. Ting, ``Panopmamba: Vision state space modeling for nuclei panoptic segmentation,'' 2026, {\em arXiv:2601.16631}.
\end{thebibliography}
@aunpublished{Kang26Panopmamba,
  author = "Ming Kang and Fung Fung Ting and Rapha{\"e}l C.-W. Phan and Zongyuan Ge and Chee-Ming Ting",
  title = "Panopmamba: Vision state space modeling for nuclei panoptic segmentation",
  howpublished = "arXiv preprint",
  note = "arXiv:2601.16631",
  year = "2026"
}
@unpublished{Kang26Panopmamba,
  author = "Kang, Ming and Ting, Fung Fung and Phan, Rapha{\"e}l C.-W. and Ge, Zongyuan and Ting, Chee-Ming",
  title = "PanopMamba: vision state space modeling for nuclei panoptic segmentation",
  note = "{a}rXiv preprint arXiv:2601.16631",
  year = "2026",
  doi= "10.48550/arXiv.2601.16631",
  url = "https://doi.org/10.48550/arXiv.2601.16631"
}
@misc{Kang26Panopmamba,
  author = "Kang, Ming and Ting, Fung Fung and Phan, Rapha{\"e}l C.-W. and Ge, Zongyuan and Ting, Chee-Ming",
  title = "{PanopMamba: Vision State Space Modeling for Nuclei Panoptic Segmentation}",
  eprint= "2601.16631",
  archivePrefix= "arXiv",
  year = "2026"
}

NOTE: Please remove some optional BibTeX fields/tags such as series, volume, address, url, and so on if the LaTeX compiler produces an error. Author names may be manually modified if not automatically abbreviated by the compiler under the control of the bibliography/reference style (i.e., .bst) file. The BibTex citation key may be bib1, b1, or ref1 when references appear in numbered style in which they are cited. The quotation mark pair "" in the field could be replaced by the brace {}, whereas the brace {} in the BibTeX field/tag title plays a role of keeping letters/characters/text original lower/uppercases or sentence/capitalized cases unchanged while using Springer Nature bibliography style files, for example, sn-nature.bst.

License

PanopMamba is released under the MIT License. Please see the LICENSE file for more information.

Copyright Notice

Many utility codes of our project base on the codes of OpenMMLab Semantic Segmentation Toolbox and Benchmark (MMSegmentation), MSVMamba, FusionMamba, MoNuSAC, CoNIC, Boundary IoU, JDTLosses, wIoU, and FWIoU repositories.