中国科学院机构知识库网格系统: 基于多个再生核希尔伯特空间的多角度人脸识别

基于多个再生核希尔伯特空间的多角度人脸识别

文献类型：期刊论文


作者	林硕; 龚志恒; 韩忠华; 史海波
刊名	光子学报
出版日期	2013
卷号	42 期号:12 页码:1436-1441
关键词	谱算法多角度预处理枝希尔伯特空间迭代
ISSN号	1004-4213
其他题名	Multi-angle Face Recognition Algorithm Based on Multi-RKHS
产权排序	1
中文摘要	针对传统谱算法在人脸识别中的局限，提出一种基于多个再生核希尔伯特空间的多角度人脸识别算法。首先，利用landmark 标记法对图像进行预处理，得到训练图像的角度，其次，通过人脸数据的多次核化迭代，使其在构建的再生枝希尔伯特空间中呈线性，针对不同类型的人脸数据，建立多个再生枝希尔伯特空间，最后，通过对比训练图像，判断待检测人脸图像的再生枝希尔伯特空间归属，实现多角反人脸识别在取FERET 和CMU-PIE 两类数据集进行对比实验，实验结果表明所提出的算法不仅在平均识别率上高于传统算法5% ，平均识别效率也较传统算法提高20%。
英文摘要	The traditional spectrum　algorithms are limited in face recognition problem. For its characteristics of problem, a novel method based on multi-reproducing Kernel Hilbert space was proposed. Firstly, the images were processed by the landmark method, and the angle of training images could be obtained. Secondly, the face data　was iterated by the　Kernel, then face data expressed linearly in the reproducing　Kernel Hilbert space. Thereafter, for many types of face data, the　multi-reproducing　Kernel　Hilbert space　were established. Finally, there producing Kernel Hilbert space belonging of human face image was judged by the comparison of training images, and the multi-angle face recognition achieved. The two classes of data sets were selected as the　experimental data, which　consisted　of　FERET　and　CMUPIE.A　large　number　of experiments were carried out. The results show that the proposed　method has great effect to recognize multi-angle face. The average recognition rate and efficiency are 5%　and 20%　higher than the traditional algorithms, respectively.
收录类别	EI ; CSCD
资助信息	国家自然科学基金(No.61164012) 和国家重大科技专项(No.2011 ZX02507-006)的资助
语种	中文
CSCD记录号	CSCD:5016011
公开日期	2013-12-26
源URL	[http://ir.sia.cn/handle/173321/13948]
专题	沈阳自动化研究所_数字工厂研究室
推荐引用方式 GB/T 7714	林硕,龚志恒,韩忠华,等. 基于多个再生核希尔伯特空间的多角度人脸识别[J]. 光子学报,2013,42(12):1436-1441.
APA	林硕,龚志恒,韩忠华,&史海波.(2013).基于多个再生核希尔伯特空间的多角度人脸识别.光子学报,42(12),1436-1441.
MLA	林硕,et al."基于多个再生核希尔伯特空间的多角度人脸识别".光子学报 42.12(2013):1436-1441.

入库方式： OAI收割

来源：沈阳自动化研究所

下载0

基于多个再生核希尔伯特空间的多角度人脸识别

其他版本