摘要
二项分布拉普拉斯边缘检测算子LoB主要利用图像的二阶微分。作者在此基础上结合图像的张量信息提出了改进的LoB滤波算子,作用于距离保持水平集方法的面积项,从而构造出新的模型。与原模型相比,该模型提高了图像的演化速度,能够分割多目标物体且对具有丰富尖角细节、深度凹陷和狭窄缝隙的图像拥有较好的分割效果。
Laplacian of binomial distribution filter （ LoB ） has been applied in image processing which mainly used second order differential of the given image. In this paper, we propose an extended LoB fiher ~ombining image structure tensor and used it into model of distance preserving level set evolution. Compared with the original model, the proposed model greatly improves the image evolution speed. Meanwhile, we perform the proposed model on some images that contain many sharp corners, deep depression and narrow gap, and get better results.
出处
《遵义师范学院学报》
2012年第6期85-88,共4页
Journal of Zunyi Normal University
基金
重庆大学研究生科技创新基金资助项目(DJXS1010025)
关键词
图像分割
二项分布拉普拉斯算子
偏微分方程
距离保持水平集方法
image segmentation
laplacian of binomial distribution filter
partial differential equation
distance preserving level set evolution
