【摘要】 针对在单样本人脸识别中每类个体的鉴别性特征难以提取的问题,提出一种基于子模式的单样本人脸识别方法.所提方法考虑了人脸的不同部位对人脸识别精度有不同的贡献度,并引入外部人脸数据集来训练学习得到每类个体的鉴别性特征.在进行人脸识别时,采用人脸校准算法提取人脸的5个基准点,并以此为中心将人脸划分成5个固定大小的子模式.在每个子模式的特征提取上,引入外部人脸数据集,并结合SVM算法训练得到属于每个子模式的分类器.最后,对每个子模式的分类结果做加权融合,得到最佳识别对象.在3个公开的人脸数据集Extend-Yale-B,ORL,AR上与现有方法进行实验比较,结果表明所提方法在识别精度上有较大提升.更多还原

【Abstract】 For the problem of difficult to extract discriminant feature of each individual in single sample face recognition,this paper proposes a single sample face recognition method based on sub-pattern.The Proposed method considers that different parts of the face have different contributions to the face recognition accuracy and discriminant feature of each individual can be learned from external face dataset.During the face recognition process,the whole face is divided into five face sub-patterns with five key points detected by the face alignment algorithm.In the feature extraction of each sub-pattern,combined with SVM algorithm to get the classifier belongs to each sub-mode with external face dataset.Finally the classification results of each person’s face sub-pattern are weighted to obtain best recognition object.The experiments were evaluated with the existing methods in Extend-Yale-B,ORL,and AR.The experimental results show that the proposed method has a great improvement in recognition accuracy.更多还原