基于图片特征与人脸姿态的人脸识别方法研究

针对人脸识别中在非限定条件下（如背景、光照等因素发生变化时）人脸多角度多姿态识别精度低的问题与现有基于识别模型的方法无法快速更新人脸类别，提出了基于图片特征与人脸姿态的识别方法，通过对人脸姿态的识别，最大程度的匹配人脸数据库中的人脸信息，使用VGG16卷积神经网络训练模型提取图片特征，生成特征向量，再使用支持向量机分别训练提取出的特征，与人脸数据库中信息进行比对，从而精确识别人脸。通过在Pubfig与FERET人脸库上实验结果表明，所采用的算法精度较高。

Research on Face Recognition Based on Image Features and Face Pose

For face recognition, the problem of low accuracy of multi-angle and multi-pose recognition in face recognition under unconstrained conditions (such as changes in background, lighting and other factors) and the existing methods based on recognition models cannot quickly update face categories, and proposes The recognition method of image features and face poses, through the recognition of face pose, maximizes the matching of face information in the face database, uses the VGG16 convolutional neural network training model to extract picture features, generate feature vectors, and then uses SVM. The extracted features are trained separately and compared with the information in the face database to accurately identify the face. Experimental results on Pubfig and FERET face databases show that the accuracy of the algorithm used is high.