基于形变模型的图像分割方法通常可以分为参数型(Parametric Deformable Models)和几何型(Geometric Deformable Models)两类.提出了一种基于几何形变理论的LBF模型.针对水平集level set模型不能处理灰度不均一图像的分割问题,采用了LBF模型,并且该模型引入了一个以高斯函数为核函数的局部二值拟合能量,以获取图像的局部信息.通过理论分析与计算机仿真算例和其他算法的性能进行对比,表明改进算法LBF提高了图像分割的稳定性和精确性,具有较高的实用价值和广泛的应用背景.%Based on the deformation model image segmentation method usually can be divided into parametric deformable models and geometric deformable models two kinds, a LBF algorithm based on region-based active contour model was presented. Aiming at overcoming the difficulties caused by intensity inhomogeneities, LBF introduce a Gaussian kernel function to define a local binary fitting energy, so that local intensity information could be embedded into a region-based active contour model. Through theory analysis and simulation examples about these algorithms , experimental results show that the LBF algorithm improves stability and accuracy of the image segmentation, which has high practical value and broad application prospect.
展开▼
