Abstract

Deep learning-based point cloud modeling has been widely investigated as an indispensable component of general shape analysis. Recently, transformer and state space model (SSM) have shown promising capacities in point cloud learning. However, limited research has been conducted on medical point clouds, which have great potential in disease diagnosis and treatment. This paper presents an SSM-based hierarchical feature learning framework for medical point cloud understanding. Specifically, we down-sample the input into multiple levels through the farthest point sampling. At each level, we perform a series of k-nearest neighbor (KNN) queries to aggregate multi-scale structural information. To assist SSM in processing point clouds, we introduce coordinate-order and inside-out scanning strategies for efficient serialization of irregular points. Point features are calculated progressively from short neighbor sequences and long point sequences through vanilla and group Point SSM blocks, to capture both local patterns and long-range dependencies. To evaluate the proposed method, we build a large-scale medical point cloud dataset named MedPointS for anatomy classification, completion, and segmentation. Extensive experiments conducted on MedPointS demonstrate that our method achieves superior performance across all tasks. The dataset is available at https://flemme-docs.readthedocs.io/en/latest/medpoints.html. Code is merged into a public medical imaging platform: https://github.com/wlsdzyzl/flemme.

Links to Paper and Supplementary Materials

Main Paper (Open Access Version): https://papers.miccai.org/miccai-2025/paper/1230_paper.pdf

SharedIt Link: Not yet available

SpringerLink (DOI): Not yet available

Supplementary Material: Not Submitted

Link to the Code Repository

https://github.com/wlsdzyzl/flemme

Link to the Dataset(s)

MedPointS dataset: https://flemme-docs.readthedocs.io/en/latest/medpoints.html

BibTex

@InProceedings{ZhaGuo_Hierarchical_MICCAI2025,
        author = { Zhang, Guoqing and Yang, Jingyun and Li, Yang},
        title = { { Hierarchical Feature Learning for Medical Point Clouds via State Space Model } },
        booktitle = {proceedings of Medical Image Computing and Computer Assisted Intervention -- MICCAI 2025},
        year = {2025},
        publisher = {Springer Nature Switzerland},
        volume = {LNCS 15969},
        month = {September},
        page = {255 -- 264}
}


Reviews

Review #1

  • Please describe the contribution of the paper

    This paper presents a State Space Model (SSM) based medical point clouds feature learning framework that employs hierarchical learning strategy to joint learn local and global point feature at multiple scales. In addition, the authors construct a large-scale medical point cloud dataset, which can benefit the research about medical point cloud. Extensive experiments on the new dataset also suggest the potential of SSMs and medical point cloud learning.

  • Please list the major strengths of the paper: you should highlight a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.
    1. The authors propose a new SSM-based hierarchical feature learning framework for medical point clouds, where coordinate-order and inside-out scanning strategies are adopted.
    2. A large-scale medical point cloud dataset is constructed for the medical point cloud-based tasks, such as classification, completion and segmentation.
    3. Comprehensive experiments are conducted for the proposed SSM-based medical point cloud feature learning framework.
  • Please list the major weaknesses of the paper. Please provide details: for instance, if you state that a formulation, way of using data, demonstration of clinical feasibility, or application is not novel, then you must provide specific references to prior work.
    1. The computation cost has not been well-studied for training and inference phases.
    2. The detail of the constructed dataset is not provided, such as the way of avoiding noise.
  • Please rate the clarity and organization of this paper

    Good

  • Please comment on the reproducibility of the paper. Please be aware that providing code and data is a plus, but not a requirement for acceptance.

    The submission has provided an anonymized link to the source code, dataset, or any other dependencies.

  • Optional: If you have any additional comments to share with the authors, please provide them here. Please also refer to our Reviewer’s guide on what makes a good review and pay specific attention to the different assessment criteria for the different paper categories: https://conferences.miccai.org/2025/en/REVIEWER-GUIDELINES.html

    N/A

  • Rate the paper on a scale of 1-6, 6 being the strongest (6-4: accept; 3-1: reject). Please use the entire range of the distribution. Spreading the score helps create a distribution for decision-making.

    (4) Weak Accept — could be accepted, dependent on rebuttal

  • Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

    1, The computation cost has not been well-studied for training and inference phases. 2, The detail of the constructed dataset is not provided, such as the way of avoiding noise.

  • Reviewer confidence

    Confident but not absolutely certain (3)

  • [Post rebuttal] After reading the authors’ rebuttal, please state your final opinion of the paper.

    N/A

  • [Post rebuttal] Please justify your final decision from above.

    N/A



Review #2

  • Please describe the contribution of the paper

    The authors propose an State Space Model (SSM) based learning framework, combining coordinate-order and inside-out scanning strategies for modeling and analyzing complex medical point cloud data. They also propose a medical point cloud dataset for evaluation of anatomical shape classification, completion and segmentation and a evaluation of the proposed method on the dataset.

  • Please list the major strengths of the paper: you should highlight a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.
    • Well written
    • Code and dataset will be available
    • The authors know the limitations of the proposed methods and they are well addressed on the paper.
    • They compare their method with similar methods showing a notable improvement.
  • Please list the major weaknesses of the paper. Please provide details: for instance, if you state that a formulation, way of using data, demonstration of clinical feasibility, or application is not novel, then you must provide specific references to prior work.
    • Small typo on page 2, line 8 -> compute vision.
  • Please rate the clarity and organization of this paper

    Good

  • Please comment on the reproducibility of the paper. Please be aware that providing code and data is a plus, but not a requirement for acceptance.

    The authors claimed to release the source code and/or dataset upon acceptance of the submission.

  • Optional: If you have any additional comments to share with the authors, please provide them here. Please also refer to our Reviewer’s guide on what makes a good review and pay specific attention to the different assessment criteria for the different paper categories: https://conferences.miccai.org/2025/en/REVIEWER-GUIDELINES.html

    N/A

  • Rate the paper on a scale of 1-6, 6 being the strongest (6-4: accept; 3-1: reject). Please use the entire range of the distribution. Spreading the score helps create a distribution for decision-making.

    (5) Accept — should be accepted, independent of rebuttal

  • Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

    The paper is well written and the method presented allows the analysis and modeling of point clouds (a format that is often used in medical imaging application like SLAM, SfM, 3D reconstruction…). Their method outperforms other existing methods on classification, completion and segmentation. The proposed dataset based on MedShapeNet (meshes) will allow the community to work on medical point cloud analysis.

  • Reviewer confidence

    Not confident (1)

  • [Post rebuttal] After reading the authors’ rebuttal, please state your final opinion of the paper.

    N/A

  • [Post rebuttal] Please justify your final decision from above.

    N/A



Review #3

  • Please describe the contribution of the paper

    The main contribution of this work is the development of an SSM-based hierarchical feature learning framework for medical point cloud understanding, which employs farthest point sampling for multi-level representation and innovative coordinate-order and inside-out scanning strategies to effectively serialize irregular point clouds. The framework utilizes vanilla and group Point SSM blocks to capture both local patterns and long-range dependencies, and is evaluated on a newly created large-scale medical point cloud dataset named MedPointS, which supports anatomy classification, completion, and segmentation tasks, with experimental results demonstrating superior performance across all these applications.

  • Please list the major strengths of the paper: you should highlight a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.

    1) it introduces a large-scale medical point cloud dataset (MedPointS), filling a significant gap in the field of medical point cloud analysis; 2) it implements and evaluates the recent Mamba/SSM architecture on this dataset, establishing a comprehensive benchmark for future research; 3) the paper is well-written and easy to follow, making its methodological contributions accessible to the research community.

  • Please list the major weaknesses of the paper. Please provide details: for instance, if you state that a formulation, way of using data, demonstration of clinical feasibility, or application is not novel, then you must provide specific references to prior work.

    The technical contribution is somewhat limited.

  • Please rate the clarity and organization of this paper

    Good

  • Please comment on the reproducibility of the paper. Please be aware that providing code and data is a plus, but not a requirement for acceptance.

    The submission has provided an anonymized link to the source code, dataset, or any other dependencies.

  • Optional: If you have any additional comments to share with the authors, please provide them here. Please also refer to our Reviewer’s guide on what makes a good review and pay specific attention to the different assessment criteria for the different paper categories: https://conferences.miccai.org/2025/en/REVIEWER-GUIDELINES.html

    N/A

  • Rate the paper on a scale of 1-6, 6 being the strongest (6-4: accept; 3-1: reject). Please use the entire range of the distribution. Spreading the score helps create a distribution for decision-making.

    (5) Accept — should be accepted, independent of rebuttal

  • Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

    A large-scale medical point cloud dataset which can fill the gap in the field of medical point cloud analysis

  • Reviewer confidence

    Very confident (4)

  • [Post rebuttal] After reading the authors’ rebuttal, please state your final opinion of the paper.

    N/A

  • [Post rebuttal] Please justify your final decision from above.

    N/A




Author Feedback

Firstly, we express our sincere appreciation for your careful and invaluable feedback. We have thoroughly examined each point raised and prepared a comprehensive response as follows:

Reviewer 1: We will conduct detailed experiments on space and time complexity analysis, which will be provided in our public code repository. Additionally, we plan to make the MedPointS dataset publicly available soon, with related documentation hosted on its dedicated dataset page.

Reviewer 2: We apologize for the typographical errors identified in the manuscript and have performed additional proofreading to address them.

Reviewer 3: This paper proposes an SSM-based feature learning framework for medical point clouds. The main technical contributions include different scanning strategies and PSSM blocks for multi-scale feature learning. Although our method shows superior performance across different tasks, the current implementation remains simple and could serve as a baseline for future research. Task-specific enhancements could be further explored in follow-up work. Furthermore, integrating language prompts and developing foundation models for medical point clouds represents a promising future direction.




Meta-Review

Meta-review #1

  • Your recommendation

    Provisional Accept

  • If your recommendation is “Provisional Reject”, then summarize the factors that went into this decision. In case you deviate from the reviewers’ recommendations, explain in detail the reasons why. You do not need to provide a justification for a recommendation of “Provisional Accept” or “Invite for Rebuttal”.

    N/A



back to top