LI Chang-xing, HUANG Yan-hu, ZHI Xiao-bin, et al. Improvements of acceleration k-means based spectral clustering algorithm for image segmentation[J]. 2016, 35(9): 137-140.
LI Chang-xing, HUANG Yan-hu, ZHI Xiao-bin, et al. Improvements of acceleration k-means based spectral clustering algorithm for image segmentation[J]. 2016, 35(9): 137-140. DOI: 10.13873/J.1000-9787(2016)09-0137-04.
基于加速k均值的谱聚类图像分割算法改进
摘要
对谱聚类图像分割算法进行改进
即引入加速均值算法替换原算法中的k均值算法
得出加速谱聚类的图像分割算法。将改进算法应用于微软剑桥研究院Grab cut数据集中的5幅实验图像
结果显示:在平均区域一致性评价不降低的前提下
改进算法完成分割所花费的平均时间比改进前可缩短58%。
Abstract
A spectral clustering algorithm for image segmentation is achieved by improving spectral clustering algorithm for image segmentation
that is
introduce acceleration k-means algorithm to replace the original k-means algorithm.The improved algorithm is applied in five experimental images from the Grab Cut dataset of Microsoft Research at Cambridge
results show that time consuming of image segmentation by the improved algorithm can be shortened by 58 % in the premise of average area of conformance assessment without reducing.