发表机构
Centro de Tecnologías de la Imagen (CTIM); Instituto Universitario de Cibernética, Empresas y Sociedad (IUCES); University of Las Palmas de Gran Canaria（图像技术中心（CTIM）; 网络、企业与社会大学研究所（IUCES）; 拉斯帕尔马斯大加那利大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究构建统一基准，在相同条件下对比3D U-Net等5种深度学习模型在BraTS 2023、2024数据集上的脑肿瘤分割性能，明确了不同架构的准确率与效率权衡及适用场景。
AI 中文摘要
从磁共振成像（MRI）中自动分割脑肿瘤已成为计算机辅助诊断、治疗规划和疾病监测的基础任务。尽管最近提出了众多深度学习架构，但客观比较仍具挑战性，因为已发表研究常采用不同的数据集、预处理策略、训练协议和评估流程。本研究提出了一个统一的实验基准，用于在相同实验条件下比较代表性卷积神经网络（CNN）、基于Transformer的模型以及最新的状态空间模型（SSM）架构。在代表不同临床场景的两个脑肿瘤分割数据集上评估了五种最先进的三维分割模型，包括3D U-Net、SegResNet、Swin UNETR、SegMamba和SegMambaV2，这两个数据集分别对应颅内脑膜瘤分割（BraTS 2023）和治疗后胶质瘤分割（BraTS 2024）。所有架构均使用相同的预处理、数据增强、优化策略和评估协议进行训练，以确保公平比较。性能通过分割准确率指标以及计算成本指标（包括推理时间和各模型的大小）进行评估。研究结果为分割准确率与计算效率之间的权衡提供了实用见解，凸显了不同架构范式在具有挑战性的三维脑肿瘤分割任务中的适用性。
英文摘要
Automatic brain tumor segmentation from magnetic resonance imaging (MRI) has become a fundamental task in computer-assisted diagnosis, treatment planning, and disease monitoring. Although numerous deep learning architectures have recently been proposed, objective comparisons remain challenging because published studies often employ different datasets, preprocessing strategies, training protocols, and evaluation procedures. This work presents a unified experimental benchmark for comparing representative convolutional neural networks (CNNs), Transformer-based models, and recent State Space Model (SSM) architectures under homogeneous experimental conditions. Five state-of-the-art three-dimensional segmentation models, including 3D U-Net, SegResNet, Swin UNETR, SegMamba, and SegMambaV2, are evaluated on two brain tumor segmentation datasets representing distinct clinical scenarios: intracranial meningioma segmentation (BraTS 2023) and post-treatment glioma segmentation (BraTS 2024). All architectures are trained using identical preprocessing, data augmentation, optimization strategies, and evaluation protocols to ensure a fair comparison. Performance is assessed using segmentation accuracy metrics together with computational cost indicators, including inference time and the size of each model. The results provide practical insights into the trade-offs between segmentation accuracy and computational efficiency, highlighting the suitability of different architectural paradigms for challenging three-dimensional brain tumor segmentation tasks.
Comments27 pages, 16 figures, 8 tables
