摘要： 提出了一种基于拉普拉斯脸和隐马尔可夫模型的视频人脸识别方法。在训练过程中，采用拉普拉斯脸方法将每一视频序列中的人脸图像映射到拉普拉斯空间，将降维后的特征作为观测值，通过隐马尔可夫模型得到每一训练视频的统计特性和时间动态特性。在识别过程中，用每一个训练视频的隐马尔可夫模型来分析测试视频的时间动态特性，计算出每一训练模型产生该序列的概率，概率最大值所对应的模型就是待识别序列所属的类别。实验结果表明，该方法能够很好地进行视频人脸识别。

关键词: 人脸识别, 位置保留映射, 拉普拉斯脸, 隐马尔可夫模型

Abstract: A method on video-based face recognition using Laplacianfaces and Hidden Markov Models(HMM) is proposed. During the training process, all the training images in each subject are projected into the obtained Laplacianspace and corresponding feature vectors, which will be used as observation vectors in the HMM training are generated. The statistics of training video sequences of each subject, and the temporal dynamics are learned by an HMM. During the recognition process, the temporal characteristics of test video sequence are analyzed over time by the HMM, and the highest score among the likelihood scores provided by HMM estimates the identity of the test video sequence. Experimental results show that the proposed method can get satisfied performance in video-based face recognition.

Key words: Face recognition, Locality preserving projection, Laplacianfaces, Hidden Markov models(HMM)