基于复合特征向量提取的隐马尔可夫实时人脸识别算法

Real-time Face Recognition Algorithm Using Hidden Markov Model Based on Complex Feature Extraction

实现了基于隐马尔可夫复合特征向量快速提取人脸识别的算法.用差分算法对实时采集到的每一帧图像快速定位到人脸区域,然后对人脸区域的数据进行规一化,并对原始图像进行DCT变换和灰度变换,以变换后的结果作为特征值对其聚类后作为隐马尔可夫模型(HMM)的观察向量,再对样本训练,训练结果制成特征脸模版存入模版库.最后通过模版对实时采集到的图像进行人脸识别.实验结果表明:该算法对复杂背景中的人脸识别具有实时性、准确性和可靠性.

An algorithm of face recognition based on Hidden Markov Model （HMM for abbreviation） is presented here. First,the difference algorithm was used to analyze each of the frames of the picture captured from camera in a real time and to position the region of the face shortly. Then,the data of the region of the face was normalized. DCT and gray transforms were applied to the raw picture, and the transformed data used as eigenvectors were clustered. Next ,the clustered eigenvectors were the vector of observation of HMM. The samples were trained and the results were saved as the face recognition template. Finally,the picture of face captured in real time was recognized by the templates. The experimental results show that the algorithm identifies the picture of face with great efficiency and Speed.