基于Snake模型的图像目标轮廓自动跟踪方法

针对实验数据像素灰度的分布特点，提出了一种对目标轮廓线进行有效和可靠的搜索和跟踪策略．由于数据中病变组织与其邻近组织像素灰度差别相对明显，首先通过来用一种改进的轮廓自动跟踪方法对目标轮廓进行跟踪，将得到的轮廓线经采样得到其离散控制点作为Snake轮廓搜索和跟踪算法的输入，既克服了Snake方法对初始轮廓线控制点分布的局限性，又避免了采用单一轮廓跟踪方法跟踪目标轮廓线的不确定性，提高了分割病变组织的速度和准确性，此方法具有较高实用意义。

Snake model based automatic tracking method for image object contour

Aiming at the distribution characteristic of pixel gray level in tested data, this paper proposes a valid and reliable searching and tracking policy for the object contour. Because of the relatively distinct contrast of gray level between unhealthy tissues and neighboring ones in data, an improved method is used to track the object contour firstly, then the contour is sampled to get the discrete controlling points of it and input them into the Snake contour tracking algorithm subroutine, thus avoiding the limitation of the initial contour controlling points of the Snake method and increasing the speed and accuracy of the segmentation for the unhealthy tissues. The proposed method is significant for practical application.