Precise segmentation of multimodal images
Precise segmentation of multimodal images
复制标题
DOI：
10.1109/tip.2005.863949
复制
发表时间：
2006-04-01
影响因子：
10.6
通讯作者：
Gimel'farb, G
中科院分区：
文献类型：
--
作者：
Farag, AA;El-Baz, AS;Gimel'farb, G
我们提出了用于多模态灰度图像无监督分割的新技术，使得每个感兴趣区域与灰度级的经验边缘概率分布的单一主导模式相关。我们遵循最传统的方法，即初始图像和所需的区域图由独立图像信号和相互依赖的区域标签的联合马尔可夫-吉布斯随机场（MGRF）模型来描述。然而，我们的重点是更准确的模型识别。为了更好地指定区域边界，图像信号的每个经验分布都通过具有正分量和负分量的高斯 (LCG) 线性组合来精确近似。我们修改了期望最大化（EM）算法来处理 LCG，并且还提出了一种新颖的基于 EM 的顺序技术来获得接近的初始 LCG 近似值，修改后的 EM 算法应从该近似值开始。所提出的技术识别混合经验分布中的各个 LCG 模型，包括正高斯和负高斯的数量。基于 LCG 模型的初始分割随后通过使用具有分析估计潜力的 MGRF 进行迭代细化。讨论了每个阶段整体分割算法的收敛性。实验表明，所开发的技术比其他已知算法更准确地分割不同类型的复杂多模态医学图像。
We propose new techniques for unsupervised segmentation of multimodal grayscale images such that each region-of-interest relates to a single dominant mode of the empirical marginal probability distribution of grey levels. We follow the most conventional approaches in that initial images and desired maps of regions are described by a joint Markov-Gibbs random field (MGRF) model of independent image signals and interdependent region labels. However, our focus is on more accurate model identification. To better specify region borders, each empirical distribution of image signals is precisely approximated by a linear combination of Gaussians (LCG) with positive and negative components. We modify an expectation-maximization (EM) algorithm to deal with the LCGs and also propose a novel EM-based sequential technique to get a close initial LCG approximation with which the modified EM algorithm should start. The proposed technique identifies individual LCG models in a mixed empirical distribution, including the number of positive and negative Gaussians. Initial segmentation based on the LCG models is then iteratively refined by using the MGRF with analytically estimated potentials. The convergence of the overall segmentation algorithm at each stage is discussed. Experiments show that the developed techniques segment different types of complex multimodal medical images more accurately than other known algorithms.