对比五种U-Net变体，发现基于Transformer的Swin UNETR在脑肿瘤和视网膜血管分割上表现最优。
From Convolution to Transformer: A Comparative Study of U-Net Variants for Brain Tumor and Retinal Vessel Segmentation
- 用Transformer替换卷积，增强长距离依赖建模能力
- Swin UNETR在脑肿瘤和视网膜数据集上分别取得0.8965和0.8078的Dice分数
- 适合需要全局上下文信息的医学图像分割任务
医学图像分割在辅助诊断、治疗规划和疾病监测中至关重要。U-Net因其编码器-解码器结构和跳跃连接被广泛用于生物医学图像分割。然而，传统的卷积型U-Net模型可能难以捕捉长距离依赖和全局上下文信息，影响复杂分割任务的表现。本文比较了五种U-Net变体：U-Net 3D、Residual U-Net、Attention U-Net、UNETR和Swin UNETR。模型在两个基准数据集上评估：BraTS 2023（脑肿瘤分割）和DRIVE（视网膜血管分割）。实验结果表明，Swin UNETR整体表现最佳，在BraTS 2023上达到0.8965的Dice分数，在DRIVE上达到0.8078。结果说明基于Transformer的U-Net变体在需要全局上下文建模的任务中更有效，而残差学习对精细结构分割仍有帮助。本研究为体积MRI和视网膜成像任务中的模型选择提供了实用参考。
原文摘要 · Abstract (English)
Medical image segmentation plays an important role in computer aided diagnosis, treatment planning, and disease monitoring. U-Net has been widely used for biomedical image segmentation because of its encoder decoder structure and skip connections. However, conventional convolution based U-Net models may have limited ability to capture long range dependencies and global contextual information, which can affect performance in complex segmentation tasks. This paper presents a comparative study of five U-Net based architectures: U-Net 3D, Residual U-Net, Attention U-Net, UNETR, and Swin UNETR. The models are evaluated on two benchmark datasets: BraTS 2023 for brain tumor segmentation and DRIVE for retinal vessel segmentation. Experimental results show that Swin UNETR achieves the best overall performance, with Dice scores of 0.8965 on BraTS 2023 and 0.8078 on DRIVE. The results suggest that transformer based U-Net variants are effective for segmentation tasks requiring global contextual modeling, while residual learning remains useful for fine structure segmentation. This study provides practical insights into model selection for medical image segmentation across volumetric MRI and retinal imaging tasks.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。