基于改进GVF snake模型的感兴趣区域轮廓提取方法

Contour Extraction Method of Region of Interest Based on Improved GVF Snake Model

针对传统模型图像边缘保持能力较差、计算量大且对弱边缘和角点的捕获能力不足等缺点,提出了一种改进的GVF snake模型。增加各向异性扩散机制,抑制传统GVF中各向同性扩散对图像边缘的模糊,并利用部分扩散、部分插值的方法进行GVF场求解,采用基于canny算子的初始轮廓线设置方法,极大地提高了图像感兴趣区域轮廓的提取效果。

Aiming at the shortcomings of the traditional model,such as poor image edge retention ability,high computational cost,and insufficient ability to capture weak edges and corners,an improved GVF snake model was proposed.Anisotropic diffusion mechanism is added to suppress the blurring of image edges caused by isotropic diffusion in traditional GVF,and the GVF field is solved by partial diffusion and partial interpolation.The initial contour setting method based on canny operator is adopted,which greatly improves the extraction effect of the contour of the region of interest in the image.