隐马尔可夫模型及在人脸识别算法中的应用

介绍了隐马尔可夫模型(HMM)及其三大算法,并将其引入人脸识别的研究中,描述了一种基于隐马尔可夫模型的人脸识别方法.一幅正面人脸图像的重要特征具有一定的顺序,它可以通过一维的HMM来建模,每个特征区域被指定为一个状态,通过K-L变换将降维以后的特征矢量作为观察矢量.和其他人脸识别的方法比较,隐马尔可夫模型更能为人脸检测和识别提供灵活的框架.

Algorithm Study for Face Recognition Using Hidden Markov Models

Introduction of the hidden Markov model( HMM) and its three algorithms, this paper combines it into a research for face recognition(FR), and describes a kind ofFR method according to HMM. The most significant facial features of a frontal face in a natural order, it may be modeled using a one-dimensional HMM by assigning each of these regions to a state, and the Karhunen Loeve(K-L) transform is used to extract a set of observation. Compared to the other methods for face recognition and detection, The HMM offers a more flexible framework for face recognition and detection.