融入全局信息的局部拟合的活动轮廓模型
A local fitting active contour model involving global information

提出一种结合局部和全局灰度信息的图像分割模型. 该模型通过引入新的控制参数，降低了局部拟合项在局部二值拟合的活动轮廓模型(LBF模型)中的主导作用，同时增加了全局信息在模型中的辅助作用，从而改进了局部二元拟合项. 实验结果表明，改进的模型不仅能分割灰度不均的图像，而且提高了轮廓初始化的灵活性. 针对一些合成和真实的图像，改进的模型减轻了对轮廓初始化位置、大小及噪声的敏感性，同时加快了曲线的演化速度，减少了CPU时间.
We propose a new image segmentation model which is combing local with global intensity of the image. It is a region-based active contour model which improves the local fitting item. With the help of controlling parameter，it can release the work of the local fitting item in local binary fitting (LBF) active contour model and enhance the weight value of the global item. The model can not only deal with the intensity inhomogeneity image segmentation problem，but also be flexible to the initialization of the contour. Experimental results show that the model has released the sensitivity of the placement and size of the contour，improved evolution speed，and decreased CPU time for synthetic and real images