用大规模脑部MRI训练3D医学图像分割，性能超越现有方法
Revisiting MAE pre-training for 3D medical image segmentation
- 基于39000个3D脑MRI构建大尺度自监督预训练数据集
- 在8个测试集上平均提升约3个Dice分数，超越nnU-Net基线
- 适配3D医学影像的残差编码器结构，适合医疗图像分割研究者
自监督学习（SSL）为海量未标注临床数据的应用提供了新可能，尤其针对标注数据稀缺的下游任务。尽管在自然语言处理和计算机视觉中取得突破，3D医学图像计算仍受限于三大问题：预训练数据集规模小、模型架构不适用于3D医学影像分析、评估方式不足。本文通过三方面改进：一、利用包含39,000个3D脑MRI体积的大规模数据集；二、在先进的nnU-Net框架中采用残差编码器U-Net结构；三、建立包含5个开发集与8个测试集的稳健评估框架，推动对3D CNN中掩码自编码器（MAE）简单概念的优化。所提出模型不仅超越以往自监督方法，还在平均上比强基线nnU-Net高出约3个Dice点，创下新基准。代码与模型已公开。
原文摘要 · Abstract (English)
Self-Supervised Learning (SSL) presents an exciting opportunity to unlock the potential of vast, untapped clinical datasets, for various downstream applications that suffer from the scarcity of labeled data. While SSL has revolutionized fields like natural language processing and computer vision, its adoption in 3D medical image computing has been limited by three key pitfalls: Small pre-training dataset sizes, architectures inadequate for 3D medical image analysis, and insufficient evaluation practices. In this paper, we address these issues by i) leveraging a large-scale dataset of 39k 3D brain MRI volumes and ii) using a Residual Encoder U-Net architecture within the state-of-the-art nnU-Net framework. iii) A robust development framework, incorporating 5 development and 8 testing brain MRI segmentation datasets, allowed performance-driven design decisions to optimize the simple concept of Masked Auto Encoders (MAEs) for 3D CNNs. The resulting model not only surpasses previous SSL methods but also outperforms the strong nnU-Net baseline by an average of approximately 3 Dice points setting a new state-of-the-art. Our code and models are made available here.
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。