一种基于改进水平集与边缘检测结合的轮廓提取与检验方法

A Contour Extraction and Verification Method Based on Improved Level Set and Edge Detection

目前图像分割方法主要研究图像分割精度、质量等,但对分割的结果研究甚少,通常是依靠人工检测结果的收敛性。通过分析图像边界的特点,给出了强边界、弱边界的定义和收敛准则,提出一种将改进的水平集与边缘检测相结合的图像分割方法,用来检验图像分割结果的收敛性,提高分割结果可信度。首先利用自适应区域生长分割出图像初始区域;然后利用预分割的结果构造初始水平集函数和进行边缘演化;最后,用边缘检测算子对分割结果进行检验,将检验得到的结果再进行水平集演化,如此往复,直到收敛。实验结果表明,该方法对弱边界图像有很好的分割效果,可较为理想地提取海马图像中的目标。

The image segmentation methods mostly are focused on image segmentation accuracy and quality etc now,but the segmentation results are rarely studied,which usually rely on personally testing the convergence of results.Through analyzing the boundary of image features,we define the strong boundary,weak boundary and convergence criterion and put forward a kind of image segmentation method combining improved level set with edge detection,which is used for testing image segmentation convergence of results and improving the segmentation result reliability.Firstly,we segment the initial area in the method of an adaptive region growing image segmentation;then,the segmentation results are utilized to construct the initial level set function and begin edge evolution later;thirdly,we use edge detection operator to inspect segmentation results,and the inspection results continue level set evolution,so back and forth,until convergence.The experimental results show that the method has good effect on weak edge image,which can ideally extract the object from Hippocampus image.