Taming Mambas for 3D Medical Image Segmentation

Recently, the field of 3D medical segmentation has been dominated by deep learning models employing Convolutional Neural Networks (CNNs) and Transformer-based architectures, each with its distinctive strengths and limitations. CNNs are constrained by a local receptive field, whereas Transformers are...

Full description

Saved in:
Bibliographic Details
Main Authors: Luca Lumetti, Vittorio Pipoli, Kevin Marchesini, Elisa Ficarra, Costantino Grana, Federico Bolelli
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11005448/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850123381820620800
author Luca Lumetti
Vittorio Pipoli
Kevin Marchesini
Elisa Ficarra
Costantino Grana
Federico Bolelli
author_facet Luca Lumetti
Vittorio Pipoli
Kevin Marchesini
Elisa Ficarra
Costantino Grana
Federico Bolelli
author_sort Luca Lumetti
collection DOAJ
description Recently, the field of 3D medical segmentation has been dominated by deep learning models employing Convolutional Neural Networks (CNNs) and Transformer-based architectures, each with its distinctive strengths and limitations. CNNs are constrained by a local receptive field, whereas Transformers are hindered by their substantial memory requirements as well as their data hunger, making them not ideal for processing 3D medical volumes at a fine-grained level. For these reasons, fully convolutional neural networks, as nnU-Net, still dominate the scene when segmenting medical structures in large 3D medical volumes. Despite numerous advancements toward developing Transformer variants with subquadratic time and memory complexity, these models still fall short in content-based reasoning. A recent breakthrough is Mamba, a Recurrent Neural Network (RNN) based on State-Space Models (SSMs), outperforming Transformers in many long-context tasks (million-length sequences) on famous natural language processing and genomic benchmarks while keeping a linear complexity. In this paper, we evaluate the effectiveness of Mamba-based architectures in comparison to state-of-the-art convolutional and Transformer-based models for 3D medical image segmentation across three well-established datasets: Synapse Abdomen, MSD BrainTumor, and ACDC. Additionally, we address the primary limitations of existing Mamba-based architectures by proposing alternative architectural designs, hence improving segmentation performances. The source code is publicly available to ensure reproducibility and facilitate further research: <uri>https://github.com/LucaLumetti/TamingMambas</uri>
format Article
id doaj-art-9bb77c0f9a50407b92e98af6735a5ad6
institution OA Journals
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-9bb77c0f9a50407b92e98af6735a5ad62025-08-20T02:34:36ZengIEEEIEEE Access2169-35362025-01-0113897488975910.1109/ACCESS.2025.357046111005448Taming Mambas for 3D Medical Image SegmentationLuca Lumetti0https://orcid.org/0000-0002-1046-332XVittorio Pipoli1https://orcid.org/0009-0008-5749-6007Kevin Marchesini2Elisa Ficarra3Costantino Grana4https://orcid.org/0000-0002-4792-2358Federico Bolelli5https://orcid.org/0000-0002-5299-6351Department of Engineering &#x201C;Enzo Ferrari,&#x201D;, University of Modena and Reggio Emilia, Modena, ItalyDepartment of Engineering &#x201C;Enzo Ferrari,&#x201D;, University of Modena and Reggio Emilia, Modena, ItalyDepartment of Engineering &#x201C;Enzo Ferrari,&#x201D;, University of Modena and Reggio Emilia, Modena, ItalyDepartment of Engineering &#x201C;Enzo Ferrari,&#x201D;, University of Modena and Reggio Emilia, Modena, ItalyDepartment of Engineering &#x201C;Enzo Ferrari,&#x201D;, University of Modena and Reggio Emilia, Modena, ItalyDepartment of Engineering &#x201C;Enzo Ferrari,&#x201D;, University of Modena and Reggio Emilia, Modena, ItalyRecently, the field of 3D medical segmentation has been dominated by deep learning models employing Convolutional Neural Networks (CNNs) and Transformer-based architectures, each with its distinctive strengths and limitations. CNNs are constrained by a local receptive field, whereas Transformers are hindered by their substantial memory requirements as well as their data hunger, making them not ideal for processing 3D medical volumes at a fine-grained level. For these reasons, fully convolutional neural networks, as nnU-Net, still dominate the scene when segmenting medical structures in large 3D medical volumes. Despite numerous advancements toward developing Transformer variants with subquadratic time and memory complexity, these models still fall short in content-based reasoning. A recent breakthrough is Mamba, a Recurrent Neural Network (RNN) based on State-Space Models (SSMs), outperforming Transformers in many long-context tasks (million-length sequences) on famous natural language processing and genomic benchmarks while keeping a linear complexity. In this paper, we evaluate the effectiveness of Mamba-based architectures in comparison to state-of-the-art convolutional and Transformer-based models for 3D medical image segmentation across three well-established datasets: Synapse Abdomen, MSD BrainTumor, and ACDC. Additionally, we address the primary limitations of existing Mamba-based architectures by proposing alternative architectural designs, hence improving segmentation performances. The source code is publicly available to ensure reproducibility and facilitate further research: <uri>https://github.com/LucaLumetti/TamingMambas</uri>https://ieeexplore.ieee.org/document/11005448/Medical imaging3D segmentationMambaU-NettransformersRNNs
spellingShingle Luca Lumetti
Vittorio Pipoli
Kevin Marchesini
Elisa Ficarra
Costantino Grana
Federico Bolelli
Taming Mambas for 3D Medical Image Segmentation
IEEE Access
Medical imaging
3D segmentation
Mamba
U-Net
transformers
RNNs
title Taming Mambas for 3D Medical Image Segmentation
title_full Taming Mambas for 3D Medical Image Segmentation
title_fullStr Taming Mambas for 3D Medical Image Segmentation
title_full_unstemmed Taming Mambas for 3D Medical Image Segmentation
title_short Taming Mambas for 3D Medical Image Segmentation
title_sort taming mambas for 3d medical image segmentation
topic Medical imaging
3D segmentation
Mamba
U-Net
transformers
RNNs
url https://ieeexplore.ieee.org/document/11005448/
work_keys_str_mv AT lucalumetti tamingmambasfor3dmedicalimagesegmentation
AT vittoriopipoli tamingmambasfor3dmedicalimagesegmentation
AT kevinmarchesini tamingmambasfor3dmedicalimagesegmentation
AT elisaficarra tamingmambasfor3dmedicalimagesegmentation
AT costantinograna tamingmambasfor3dmedicalimagesegmentation
AT federicobolelli tamingmambasfor3dmedicalimagesegmentation