利用PET图像进行诊治时需要对人体病灶精确定位，PET图像中病灶目标区域的分割是早期诊断与治疗的前提和关键。基于传统Snake模型的方法在PET图像分割时存在对初始轮廓过于敏感，难以收敛到目标凹型区域等问题，为此将GVF Snake模型引入PET图像的分割中。为防止GVF Snake模型陷入局部最优，进一步利用差分进化(DE)算法的全局优化特性对GVF Snake模型分割的结果进行优化，提高PET图像分割精度。实验结果表明，该方法能有效地对PET图像中的病灶目标区域进行分割，可避免陷入局部最优且具有良好的实时性。

Focus contour extraction of PET images is of immense significance for the treatment of malignant tumors and cardiovascular diseases etc. However, the commonly used method based on the traditional snake model is sensitive to the position of the initial curve and it’s hard to converge to the concave boundary of the object. To address these problems, an improved GVF Snake model based on DE algorithm is proposed in this paper. The results from comparative experiments of extracting contour of human brain demonstrate that the new model is an effective method for segmenting the PET images.