Abstract: As the Geodesic Active Contour（GAC） model brings boundary leakage easily and sensitive to the initialization, and Local Image Fitting（LIF） model traps into local minimums easily, a level set model based on edge and region information is proposed, which can automatically adjust the proportion of GAC and LIF in the fusing model by setting weight. Experiment results demonstrate the method is superior to GAC and LIF model both in terms of speed and effect.

Key words: Geodesic Active Contour（GAC） model, Local Image Fitting(LIF) model, level set algorithm, intensity inhomogeneity

摘要： 针对测地线主动轮廓（GAC）模型容易产生边界泄露且对初始位置敏感及局部图像拟合（LIF）模型容易陷入局部极小的问题，提出融合边缘与区域模型的水平集算法。通过设置权值，该算法能自适应地调整GAC模型和LIF模型在融合算法中所占的比例。对不同图像的实验结果表明该算法的迭代收敛速度比GAC模型和LIF模型要快，分割效果明显优于GAC模型和LIF模型。

关键词: 测地线主动轮廓（GAC）模型, 局部图像拟合（LIF）模型, 水平集算法, 灰度不均匀