[论文解读] SAM vs BET: A Comparative Study for Brain Extraction and Segmentation of Magnetic Resonance Images using Deep Learning

Sovesh Mohapatra, Advait Gosai+1 more|arXiv (Cornell University)|Apr 10, 2023

Medical Image Segmentation Techniques被引用 20

一句话总结

该论文将 Segment Anything Model (SAM) 与 FSL BET 在多样化 MR 图像上的脑提取进行比较，显示 SAM 在 Dice、IoU 与准确率方面通常优于 BET，尤其是在病变和低质图像情况下，同时也展示了 SAM 在脑内结构分割方面的潜力。

ABSTRACT

Brain extraction is a critical preprocessing step in various neuroimaging studies, particularly enabling accurate separation of brain from non-brain tissue and segmentation of relevant within-brain tissue compartments and structures using Magnetic Resonance Imaging (MRI) data. FSL's Brain Extraction Tool (BET), although considered the current gold standard for automatic brain extraction, presents limitations and can lead to errors such as over-extraction in brains with lesions affecting the outer parts of the brain, inaccurate differentiation between brain tissue and surrounding meninges, and susceptibility to image quality issues. Recent advances in computer vision research have led to the development of the Segment Anything Model (SAM) by Meta AI, which has demonstrated remarkable potential in zero-shot segmentation of objects in real-world scenarios. In the current paper, we present a comparative analysis of brain extraction techniques comparing SAM with a widely used and current gold standard technique called BET on a variety of brain scans with varying image qualities, MR sequences, and brain lesions affecting different brain regions. We find that SAM outperforms BET based on average Dice coefficient, IoU and accuracy metrics, particularly in cases where image quality is compromised by signal inhomogeneities, non-isotropic voxel resolutions, or the presence of brain lesions that are located near (or involve) the outer regions of the brain and the meninges. In addition, SAM has also unsurpassed segmentation properties allowing a fine grain separation of different issue compartments and different brain structures. These results suggest that SAM has the potential to emerge as a more accurate, robust and versatile tool for a broad range of brain extraction and segmentation applications.

研究动机与目标

• 评估 SAM 相对于 BET 在跨越多样化 MRI 数据集的自动脑提取中的性能。
• 评估 SAM 超越全脑提取、对特定脑区和病变的分割能力。
• 识别在图像质量变化和病变存在等情形下，SAM 相对于 BET 提供鲁棒性优势的场景。
• 讨论预处理管线的实际影响及 SAM 的潜在局限性。

提出的方法

• 使用来自 ATLAS、WMH Challenge 及内部数据集的 45 张匿名化 MR 大脑图像，覆盖五个类别。
• 将所有图像标准化到 MNI152 空间以实现公平比较。
• 使用 Dice、IoU、Accuracy、Recall 和 Precision 等指标比较 BET 和 SAM 的脑提取。
• 对于 SAM，采用带自定义边界框算法的二维切片工作流，以决定包括/排除区域并生成掩模。
• 将 3D MRI 数据转换为 SAM 输入的二维切片，并重建三维脑提取输出。
• 在平面和模态上提供定性可视比较。

Figure 1: Comprehensive SAM workflow – visualizing the end-to-end process with custom box algorithm, highlighting blue dots (exclusion mask) and yellow dots (inclusion mask)

实验结果

研究问题

• RQ1 SAM 是否在多样化的 MRI 模态和病理条件下提供比 BET 更高的 Dice、IoU 和准确性？
• RQ2 图像质量、归一化以及病变的存在如何影响 SAM 和 BET 的性能？
• RQ3 SAM 是否能够可靠地分割脑内结构（例如胼胝体、基底节）以及超越全脑提取的病变？
• RQ4 在神经影像学管线中使用 SAM 进行脑提取时，实际权衡有哪些（如计算要求）？

主要发现

• SAM 在大多数模态和数据集上通常在 Dice、IoU 和 Accuracy 上优于 BET。
• BET 在高质量、归一化的图像中相对表现更好，但在脑边界附近的病变和较低质量数据方面表现欠佳。
• SAM 的自定义边界框方法能够更准确地包含/排除区域，有助于近脑外区域的结构与病变分割。
• 在 WMH 3DT1 数据集中，SAM 达到 Dice 0.914 和 IoU 0.842，而 BET 的 Dice 0.628 和 IoU 0.518（有引用的示例）。
• SAM 在某些 FLAIR WMH 情况下可能过度裁剪，但在多模态下仍显示出稳健的性能；总体而言，SAM 展示了更强的鲁棒性和用于脑提取及区域分割的多样性。
• 基于 SAM 的工作流需要将 3D 数据转换为 2D 切片，相较于 BET 可能需要更长的处理时间（约 ~30 秒总计）。

Figure 2: Comparison of original MRI images and extraction outputs by BET and SAM for A. sagittal, B. coronal, and C. axial anatomical planes. BET extraction results in sagittal, coronal, and axial planes, respectively. SAM extraction results in sagittal, coronal, and axial planes, respectively. The