Rohlfing Torsten, Russakoff Daniel B, Maurer Calvin R
Image Guidance Laboratories, Department of Neurosurgery, Stanford University, Stanford, CA, USA.
Inf Process Med Imaging. 2003 Jul;18:210-21. doi: 10.1007/978-3-540-45087-0_18.
It is well-known in the pattern recognition community that the accuracy of classifications obtained by combining decisions made by independent classifiers can be substantially higher that the accuracy of the individual classifiers. In order to combine multiple segmentations we introduce two extensions to an expectation maximization (EM) algorithm for ground truth estimation based on multiple experts (Warfield et al., MICCAI 2002). The first method repeatedly applies the Warfield algorithm with a subsequent integration step. The second method is a multi-label extension of the Warfield algorithm. Both extensions integrate multiple segmentations into one that is closer to the unknown ground truth than the individual segmentations. In atlas-based image segmentation, multiple classifiers arise naturally by applying different registration methods to the same atlas, or the same registration method to different atlases, or both. We perform a validation study designed to quantify the success of classifier combination methods in atlas-based segmentation. By applying random deformations, a given ground truth atlas is transformed into multiple segmentations that could result from imperfect registrations of an image to multiple atlas images. We demonstrate that a segmentation produced by combining multiple individual registration-based segmentations is more accurate for the two EM methods we propose than for simple label averaging.
在模式识别领域众所周知，通过组合独立分类器所做的决策而获得的分类准确率，可能会显著高于单个分类器的准确率。为了组合多个分割结果，我们对基于多位专家的期望最大化（EM）算法（Warfield等人，MICCAI 2002）进行了两种扩展，用于估计真实情况。第一种方法是重复应用Warfield算法并随后进行整合步骤。第二种方法是Warfield算法的多标签扩展。这两种扩展都将多个分割结果整合为一个比单个分割结果更接近未知真实情况的分割结果。在基于图谱的图像分割中，通过对同一图谱应用不同的配准方法，或对不同图谱应用相同的配准方法，或两者兼用，自然会产生多个分类器。我们进行了一项验证研究，旨在量化基于图谱的分割中分类器组合方法的成功程度。通过应用随机变形，将给定的真实图谱转换为多个分割结果，这些分割结果可能是由于图像与多个图谱图像的不完美配准而产生的。我们证明，对于我们提出的两种EM方法，通过组合多个基于个体配准的分割结果所产生的分割，比简单的标签平均更准确。
