摘要
人脸识别是生物特征识别领域的一个重要研究内容,并因为具有深厚的学术背景和广泛的应用前景从而被大学者关注。根据三维人脸不变特性描述的数据特点,引入了区域稀疏表示和低秩矩阵恢复理论。通过谱回归和矩阵完整性约束,从带噪声的原始数据中得到干净的输入数据,获得鲁棒的低秩投影矩阵,并将低维特征矢量应用在三维人脸识别中。实验结果说明本文提出的三维人脸特征提取算法具有更多的辨别性、鲁棒性和通用性,具有良好的三维数据表达能力。
Face recognition is an important research content in the field of biometric recognition,and it has been concerned by universities because of its profound academic background and wide application prospect. According to the data characteristics of 3D face invariant features,the region sparse representation and low rank matrix recovery theory are introduced. By regression and matrix spectral integrity constraints derived from the raw data in a clean noisy input data,obtained robust low rank projection matrix,and low-dimensional feature vector used in the three-dimensional face recognition. The experimental results show that the proposed 3D face feature extraction algorithm has more discrimination,robustness and generality,and has a good ability to express 3D data.
出处
《激光杂志》
北大核心
2015年第11期67-70,共4页
Laser Journal
基金
2015年河南省科技计划软科学项目(152400410598)
河南省教育厅科学技术研究重点项目(15A520055)
关键词
图像特征提取
区域稀疏
矩阵
三维数据
image feature extraction
region sparse
matrix
three-dimensional data
