Abstract: The active contour model based on global information can not effectively segment images with intensity inhomogeneities. While, the active contour model based on local information is sensitive to the location of initial contour. Therefore, a new Signed Pressure Force（SPF） function is constructed, which combins global information with local information. And the SPF of Selective Binary and Gaussian Filtering Regularized Level Set（SBGFRLS） model is replaced by the new SPF. Meanwhile, a new balloon force function is constructed. Level set method of SBGFRLS in evolving contour curve is continued to use. Experimental results demonstrate the high accuracy of the segmentation results on images with intensity inhomogeneities, less sensitive to the location of initial contour, and good anti-noise performance.

Key words: active contour model, image segmentation, signed pressure force function, level set

摘要： 基于全局信息的活动轮廓模型不能有效分割灰度不均匀图像，而基于局部信息的活动轮廓模型对轮廓初始化位置比较敏感。为此，提出结合全局信息和局部信息，构造新的符号压力函数（Signed Pressure Force，SPF），替代Selective Binary and Gaussian Filtering Regularized Level Set（SBGFRLS）模型中的符号压力函数，同时构造一种新的气球力函数，并采用SBGFRLS水平集方法演化轮廓曲线来分割图像的方法。实验结果证明该方法能有效分割灰度不均图像，同时对轮廓初始化位置不敏感，对噪声有较好的抗干扰性。

关键词: 活动轮廓模型, 图像分割, 符号压力函数, 水平集