On the application of the enhanced Hoshen-Kopelman algorithm for image analysis

被引：15
作者：
Hoshen, J
机构：
[1] Naperville, IL 60540-7112
来源：
PATTERN RECOGNITION LETTERS | 1998年 / 19卷 / 07期
关键词：
cluster analysis; connected components; percolation; spatial clusters;
D O I：
10.1016/S0167-8655(98)00018-X
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
The enhanced Hoshen-Kopelman (EHK) algorithm is introduced for a single pass analysis of spatial clusters in large images. The EHK algorithm is a generalisation of the algorithm known in the statistical physics literature as the Hoshen-Kopelman (HK) algorithm. While the HK algorithm was designed to compute cluster sizes in a binary image, the EHK algorithm enables the computation of cluster shape parameters such as spatial clusters moments, perimeters and bounding boxes in a multiple class image. An example of spatial cluster analysis for a simulated image containing 2 x 10(9) pixels is given. (C) 1998 Elsevier Science B.V. All rights reserved.
引用
收藏
页码：575 / 584
页数：10
相关论文
共 25 条
123