一种改进的基于蛇模型的分割方法

An Optimized Segmentation Method Based on Snake Model

CT图像的分割在临床的诊断和治疗中有着重大的意义。其中Snake分割算法能够得到较好的分割结果,该文通过对传统Snake算法基本原理的研究,提出了一种改进的Snake分割模型。首先,通过数学形态学的方法得到CT图像的边缘,其次,运用改进的能量方程对分割过程进行迭代。优化的Snake模型能够克服传统Snake模型无法收敛于极凹处以及对噪声敏感的问题。通过实验将模型应用于实际CT图像分割,并且得到较精确的实验结果。

Segmentation of the CT images is a meaningful step in diagnose of clinical injuries and treatment. Snake model can get a better result among varieties of segmentation methods. An optimized segmentation method based on snake model （or paramet- ric active contour model） is proposed. Firstly, an edge map is generated by morphology operation from the original CT images. Then, the proposed function is used to iterate in the segmentation process. The model addresses the problems of unable to con- verge to concavity and noise sensitivity of the traditional snake models. And the model can be applied to practical usage with ac- curate results.