基于对偶方法和局部统计信息的快速图像分割

CV模型和局部二值拟合模型用于图像分割时往往只能得到局部最优解，且计算量大，分割速度慢。为此，引入一个变量，将其与已知变量组成对偶变量，并利用图像的局部统计信息，建立主动轮廓模型的对偶模型，实现图像的快速分割。针对合成图像、多目标物体图像和灰度不均匀的医学图像进行实验，结果表明，该模型能自动处理拓扑结构的变化，从而快速准确地分割图像。

Fast Image Segmentation Based on Dual Method and Local Statistic Information

Chan-Vese（CV） model and Local Binary Fitting（LBF） model always have local optimal solutions when they are used in image segmentation, and they are very computationally expensive and segmentation speed is slow. This paper introduces a variable to become dual variable by means of known variable and uses local statistic information to establish the dual model of active contour model, so that the segmentation precision and speed are improved. Experiments are made on synthetic images, multi-target object images and medical images with intensity inhomogeneity, whose results prove that the proposed model can automatically handle topological changes and get precise segmentation results.