流形学习及其在检索中的应用
首发时间:2011-08-19
摘要:作为一种新的非监督性统计学习方法,流形学习近年来越来越引起机器学习及认知科学工作者的重视。其本质是发现高维观测数据集的内在低维流形结构和嵌入映射关系。流形学习的各种算法也成为研究热点,很多经典流形学习算法包括多维尺度分析(Multidimensional Scaling, MDS)、主成分分析(Principle Component Analysis, PCA)、拉普拉斯特征变换(Laplacian Eigenmaps, LE)、局部线性嵌入(Locally Linear Embedding, LLE)、等度规映射(Isometric Mapping, Isomap)等在图像信息压缩、模式识别、图像处理等众多领域都有广泛应用。本文主要介绍了流形学习的研究背景,各个算法的数学描述,在图像视频检索领域的应用以及未来的研究方向。
For information in English, please click here
Manifold Learning and Its Application in Retrieval
Abstract:As a new unsupervised learning method, manifold learning has captured the attention of researchers in the field of machine learning and cognitive science. Its nature is to find the relation between intrinsic low-dimensional manifold structure and embedded mapping in the high-dimensional observed data set. Many algorithms in manifold learning have been study hotspot and widely applied in image information compression, pattern recognition and image processing. This paper mainly illustrates the research background, each algorithm's description in math, the application in image and video retrieval, and the future study direction of manifold learning.
Keywords: manifold learning image retrieval video retrieval compressive sensing
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
流形学习及其在检索中的应用
评论
全部评论0/1000