融合局部和全局高斯概率信息的图像分割模型

在现有的活动轮廓中，LBF模型、LIF模型和LGDF模型是著名的基于区域的模型。虽然能分割灰度不均匀的图像，但对活动轮廓的初始化和噪声较为敏感。针对该问题，提出一种融合全高斯和局部高斯概率信息的活动轮廓模型。首先由全局高斯模型的全局灰度拟合力和局部高斯模型的局部灰度拟合力的一个线性组合来构造水平集演化力，然后引入这两个拟合力的动态权重以达到该模型的灵活性，实验结果表明，该模型能分割灰度不均的图像，且允许灵活的轮廓初始化，抗噪声性强。

Image segmentation model driven by local and global Gauss- ian probability information

In the existing active contour models, Local Binary Fitting（LBF）model、Local Image Fitting（LIF）model and Local Gaussian Distribution Fitting（LGDF）model are popular region-based models. The three models are able to deal with intensity inhomogeneity, however, they are sensitive to initialization and noise. In order to address this problem, an active contour model combining the global and local Gaussian probability information is proposed. First the force of level set evolution is defined as a linear combination of the global and local Gaussian distribution model. Then the dynamic weights of the forces are introduced to allow the flexible active contour model. Experimental results show the the proposed model can segment images with intensity inhomogeneity, while it allows flexible initialization and is less sensitive to noise.