Multiresolution segmentation of natural images: From linear to nonlinear scale-space representations
Multiresolution segmentation of natural images: From linear to nonlinear scale-space representations
复制标题
DOI：
10.1109/tip.2004.828431
复制
发表时间：
2004-08-01
影响因子：
10.6
通讯作者：
Vandergheynst, P
中科院分区：
文献类型：
--
作者：
Petrovic, A;Escoda, OD;Vandergheynst, P
在本文中，我们介绍了一个框架，合并了经典的思想借用尺度空间和多分辨率分割与非线性偏微分方程。一个非线性尺度空间堆栈构造通过一个适当的扩散方程。该堆栈进行了分析，并构建了一个树的相干段之间的关系的基础上，不同的尺度层。修剪该树被证明是用于不同类别图像的无监督分割的非常有效的工具（例如，天然的、医学的等等）。这种技术是轻的计算的观点，可以扩展到非标量数据在一个简单的方式。
In this paper, we introduce a framework that merges classical ideas borrowed from scale-space and multiresolution segmentation with nonlinear partial differential equations. A nonlinear scale-space stack is constructed by means of an appropriate diffusion equation. This stack is analyzed and a tree of coherent segments is constructed based on relationships between different scale layers. Pruning this tree proves to be a very efficient tool for unsupervised segmentation of different classes of images (e.g., natural, medical, etc.). This technique is light on the computational point of view and can be extended to nonscalar data in a straightforward manner.