manifolds

  流形学习是近年来机器学习及模式识别等领域的一个研究热点，其主要目标是去发现高维观察数据空间的低维光滑流形。自从2000年Roweis和Saul提出LLE算法、Tenenbaum等人提出Isomap算法，特别是Donoho等人发现Isomap算法能够准确发现人脸图像流形潜在的参数空间、张长水等人将LLE算法用于人脸识别并取得了较好的识别效果之后，基于流形学习的人脸识别研究引起了人们的广泛关注。(Manifold learning in recent years the field of machine learning and pattern recognition as a research focus, its main goal is to find high-dimensional data space observation of low-dimensional smooth manifold. Since 2000 LLE algorithm proposed by Roweis and Saul, Tenenbaum and others proposed Isomap algorithm, especially Donoho, who found that Isomap algorithm can find the potential face images of the parameter space manifold, and Chang-water will LLE algorithm for face recognition and to obtain a good recognition results, the manifold learning-based face recognition research has aroused extensive attention.)