摘要： SLGS模型不能处理灰度不均图像，而RSF模型对初始轮廓较敏感。为此，将RSF模型定义的局部信息融入SLGS模型中，提出一个以偏微分方程形式表达的活动轮廓模型。利用SLGS模型全局拟合量和RSF模型局部拟合量的线性组合构造符号压力函数，调整拟合量的权重以提升模型对灰度不均图像的处理能力和轮廓初始化的灵活性，并利用高斯滤波正则水平集函数法实现水平集函数的正则化。实验结果表明，该模型的分割结果比SLGS模型更准确。

关键词: 图像分割, 水平集方法, 活动轮廓模型, C-V模型, RSF模型, SLGS模型, 灰度不均

Abstract: The SLGS model can not deal with intensity inhomogeneity, while the RSF model is sensitive to initialization. Incorporating the local information defined by the RSF model into the SLGS model, this paper proposes a novel active contour model in a partial differential equation formulation. A signed pressure force function is defined by a linear combination of the global fitting values defined by the SLGS model and the local fitting values defined by the RSF model. By choosing appropriately the parameter that controls the influence of the global and local fitting values, the proposed model can efficiently deal with intensity inhomogeneity and allows for the flexible initialization. Experimental results show that the proposed model is more accurate than the SLGS model.

Key words: image segmentation, level set method, active contour model, C-V model, RSF model, SLGS model, intensity inhomogeneity

中图分类号: 

• TP911.73