中文摘要：

针对传统Snake模型在图像目标检测和分割时不能处理拓扑变化以及不能反映出演化曲线的内在几何特性，提出了一种改进模型——基于水平集的测地主动轮廓模型。该模型采用测地主动轮廓模型，并结合水平集方法，即用水平集函数表示测地主动轮廓模型的曲线演化方程，来模拟初始曲线沿能量下降最快的方向演化的过程。对该模型进行研究，将其应用于细胞图像的目标检测和分割实验中，实验结果表明，所提出的新方法具有良好的检测效果，对多目标进行了有效分割，并且能清晰反映出演化曲线的内在几何特性以及具有良好的拓扑处理能力，这些特性是传统Snake模型所不具有的。

英文摘要：

Aimed at the original Snake model can＇t handle topological changes and represent clear geometric quantities in object detecting and segmenting, a novel model for object detecting and segmenting was presented. Based on geodesic active contour model, the new approach, which adopted level set function to express the model curve evolution equation, to analog the dynamic curve evolves in the direction of energy reduces mostly. Research on the approach and applied it to the object detecting and segmenting experiment with some cell pictures, the results showed that the proposed new approach obtained very good effect and was much more powerful than other traditional approaches, since it was able to handle topological changes and represent clear geometric quantities of the dynamic curve.