Neural Memory State Space Models for Medical Image Segmentation
International Journal of Neural Systems · 2025-01 · 35 (01)
DOI: 10.1142/s0129065724500680 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Neural Systems, online erschienen am 30.09.2024. Verfasst von Zhihua Wang und 8 weiteren Autor:innen. Beteiligte Einrichtung: College of Computer Science, Zhejiang University, Hangzhou, P. R. China. Hauptthemen laut MEDLINE: Deep Learning, Neural Networks, Computer. Die Arbeit wurde bisher 11-mal zitiert (Crossref) und nennt 36 Literaturangaben. Gefördert von: Joint Fund for Medical Artificial Intelligence, Anhui Provincial Key Research and Development Project, National Natural Science Foundation of China.
Abstract (Originalsprache)
With the rapid advancement of deep learning, computer-aided diagnosis and treatment have become crucial in medicine. UNet is a widely used architecture for medical image segmentation, and various methods for improving UNet have been extensively explored. One popular approach is incorporating transformers, though their quadratic computational complexity poses challenges. Recently, State-Space Models (SSMs), exemplified by Mamba, have gained significant attention as a promising alternative due to their linear computational complexity. Another approach, neural memory Ordinary Differential Equations (nmODEs), exhibits similar principles and achieves good results. In this paper, we explore the respective strengths and weaknesses of nmODEs and SSMs and propose a novel architecture, the nmSSM decoder, which combines the advantages of both approaches. This architecture possesses powerful nonlinear representation capabilities while retaining the ability to preserve input and process global information. We construct nmSSM-UNet using the nmSSM decoder and conduct comprehensive experiments on the PH2, ISIC2018, and BU-COCO datasets to validate its effectiveness in medical image segmentation. The results demonstrate the promising application value of nmSSM-UNet. Additionally, we conducted ablation experiments to verify the effectiveness of our proposed improvements on SSMs and nmODEs.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- College of Computer Science, Zhejiang University, Hangzhou, P. R. China.
- Zhejiang Provincial Key Laboratory of Service Robot, Hangzhou, Zhejiang Province, P. R. China.
- Department of Ultrasound, The First Affiliated Hospital of Anhui Medical University, Hefei, P. R. China.
- College of Computer Science, Sichuan University, Chengdu, P. R. China.
- Department of Cardiovascular Surgery, The Second Xiangya Hospital, Central South University, Changsha, P. R. China.
- Department of Ophthalmology, Eye Center, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei, P. R. China.
ORCID-Profile: Zhihua Wang, Wang Zhou, Quansong He, Li Lu, Tao He, Jiajun Bu
Förderung
- Joint Fund for Medical Artificial Intelligence
- Anhui Provincial Key Research and Development Project
- National Natural Science Foundation of China
- China Postdoctoral Science Foundation
Zitieren
APA
Wang, Z., Gu, J., Zhou, W., He, Q., Zhao, T., Guo, J., Lu, L., He, T., & Bu, J. (2025). Neural Memory State Space Models for Medical Image Segmentation. International Journal of Neural Systems, 35(01). https://doi.org/10.1142/s0129065724500680
Vancouver
Wang Z, Gu J, Zhou W, He Q, Zhao T, Guo J, et al. Neural Memory State Space Models for Medical Image Segmentation. International Journal of Neural Systems. 2025;35(01). doi:10.1142/s0129065724500680
BibTeX
@article{Wang2025,
author = {Wang, Zhihua and Gu, Jingjun and Zhou, Wang and He, Quansong and Zhao, Tianli and Guo, Jialong and Lu, Li and He, Tao and Bu, Jiajun},
title = {Neural Memory State Space Models for Medical Image Segmentation},
journal = {International Journal of Neural Systems},
year = {2025},
volume = {35},
number = {01},
doi = {10.1142/s0129065724500680}
}
EUVIMED