摘要
文章实现了一个基于内容的图像检索系统.该系统引入了基于非负矩阵分解NMF(Non-nega-tive M atrix Factorization)相关反馈技术及局部线性嵌入(Locally L inear Embedd ing)图像特征降维方法.在由500幅图像组成的图像库上进行的实验,结果证明了该系统具有较好的查准率.
This paper implements a content - based image retrieval system. The system introduces NMF - based relevance feedback and Locally Linear Embedding in dimensionality reduction. Experiments were carried out on a database of 500 images. It shows that more images relevant to the query image can be found efficiently.
出处
《宜春学院学报》
2006年第6期22-25,共4页
Journal of Yichun University
基金
浙江省教育厅资助项目(编号:20060495)
关键词
图像检索
非负矩阵分解
局部线性嵌入
Image Retrieval
Non- negative Matrix Factorization
Locally Linear Embedding
