摘要
针对不同视角的行人样本具有较大的类内差异性,造成多视角行人识别错误率较高的问题,提出一种基于非负矩阵分解最小二乘的多视角行人分类算法.采用非负矩阵分解的方法对多视角的行人样本图像进行子空间分解,提取基向量;引入协同表示的方法并在最小二乘约束下,对子空间进行稀疏表示获得稀疏分解系数;利用近邻子空间方法对分解系数进行分类.基于自行构建的多视角行人数据库进行对比实验,结果表明该算法的准确性和有效性优于其他方法.
Multi-view pedestrian samples have so high intra-class variance that multi-view pedes-trian classification suffers from high classification error.To solve this problem,a novel multi-view pedestrian recognition algorithm is proposed in this paper based on non-negative matrix fac-torization (NMF)and least square.Firstly,by using the NMF,the subspace of multi-view pedes-trian samples is acquired and base vectors are extracted.Secondly,by introducing the collabora-tive representation the sparse presentation of the subspace is performed and then constrained by the least square,sparse coefficients are obtained.Finally,multi-viewpoint classification is com-pleted by using sparse coefficients based on the nearest subspace rule.The comparison experi-mental results on the self-established multi-view pedestrian dataset show that the proposed meth-od outperforms several state-of-the-art methods in terms of accuracy and effectiveness.
出处
《陕西师范大学学报(自然科学版)》
CAS
CSCD
北大核心
2014年第4期10-15,共6页
Journal of Shaanxi Normal University:Natural Science Edition
基金
国家自然科学基金资助项目(61303186)
国防科技大学优秀学位论文选题资助项目(43451332142)
关键词
非负矩阵分解
非负最小二乘
稀疏表示
多视角分类
non-negative matrix factorization
non-negative least square
sparse representation
multi-view classification