主动轮廓模型的一种初始轮廓设定方法

An Initial Contour Method for Active Contour Model

图像分割是图像处理中至关重要的一步,是进行图像分析与图像理解的基础,分割效果的好坏直接关系到后续的图像处理。针对目前主动轮廓模型在图像分割领域中分割速度慢的缺点,提出了一种新的初始轮廓设置方法。首先将图像进行预处理得到一个或几个较粗糙的连续轮廓;然后利用八邻域分割法,检测出该边缘作为梯度矢量流主动轮廓模型的初始轮廓,经过迭代后得到较精确的收敛轮廓。大量医学图像分割实验表明,运用本文提出的方法能够较好地检测到目标边界,速度快,是一种较为理想的图像分割方法。

Image segmentation which is the basis of image analysis and image understanding is a very important step in the area of image processing. The result of image segmentation determines whether it is suitable to go on the following processing. Considering the defect of the existing image segmentation method such as the slow speed of segmentation and acquiring more manual operations in the area of medical image segmentation, a new method of setting initial contour is proposed in this paper. Firstly the image will be processed simply in which a rough and continuous contour will be obtained, then 8-adjacent segmentation method will be adopted to detect the edge, finally the initial contour will converge to the boundary needed by using gradient vector flow field. Experimental results show that, this method can get a better result of convergence of the target boundary, and is an ideal method for medical image segmentation with fast speed and high degree of automation