BriefGPT.xyz
Sep, 2013
使用二阶非负矩阵分解进行高光谱图像的分层聚类
Hierarchical Clustering of Hyperspectral Images using Rank-Two Nonnegative Matrix Factorization
HTML
PDF
Nicolas Gillis, Da Kuang, Haesun Park
TL;DR
本文介绍了一种针对高分辨率高光谱图像的层次聚类算法,该算法采用新型的二阶非负矩阵分解,能够高效、稳定地将像素划分为多个簇,并可用于提取纯点。实验结果表明,与k-means、球形k-means和标准NMF等标准聚类技术相比,该方法的性能更好。
Abstract
In this paper, we use two simple tools, namely,
convex geometry
and rank-two nonnegative matrix factorizations, to design a
hierarchical clustering
algorithm for high-resolution
→