采用多任务稳健主成分分析的运动目标分割(英文)

被引：1

作者：

王向阳 ^{[1
,2
]}

万旺根 ^{[1
,2
]}

机构：

[1] 上海大学通信与信息工程学院

[2] 上海大学智慧城市研究院

来源：

应用科学学报 | 2014年 / 32卷 / 05期

关键词：

运动分割; 低秩矩阵恢复; 稀疏表示; 稳健主成分分析; 增广拉格朗日乘子法;

D O I：

暂无

中图分类号：

TP391.41 [];

学科分类号：

080203 ;

摘要：

提出一种多任务稳健主成分分析方法,用以结合多视觉特征实现运动目标分割.给定由多类型特征矩阵描述的视频数据,将它分解为低秩和稀疏部分,其中的稀疏部分对应于运动目标.该矩阵分解过程是一个凸优化问题,通过用ALM方法最小化核范数和2,1-范数的约束组合.与仅利用单类型特征的方法相比,本文提出的方法能够结合多类型特征,因此可获得更加精确可靠的结果.对HumanEva和Change Detection两个数据集的实验表明了该方法的有效性.

引用

页码：473 / 480

页数：8

共 13 条

[1]

Saliency detection by multitask sparsity pursuit. Lang, Congyan,Liu, Guangcan,Yu, Jian,Yan, Shuicheng. IEEE Transactions on Image Processing . 2012

[2]

A Singular Value Thresholding Algorithm for Matrix Completion. Jian-Feng Cai,Emmanuel J. Candès,Zuowei Shen. SIAM Journal on Optimization . 2010

[3]

Changedetection.net:a new change detection benchmark dataset. Goyette N,Jodoin P M,Porikli F,Konrad J,Ishwar P. IEEE Workshop on Change Detection (CDW-12)at CVPR 2012 . 2012

[4]

Robust principal component analysis for background subtraction:systematic evaluation and comparative analysis. Guyon C,Bouwmans T,Zahzah E. Principal Component Analysis . 2012

[5]

Multi-task low-rank affinities pursuit forimage segmentation. B. Cheng,G. Liu,Z. Huang,S. Yan. IEEE International Conference on Computer Vision . 2011

[6]

Robust principal component analy-sis:exact recovery of corrupted Low-rank matrices by convex op-timization. Wright J,Peng Y G,Ma Y. Proceedings of Advances in Neural InformationProcessing Systems . 2009

[7]

Bayesian robust principal component analysis. Ding, Xinghao,He, Lihan,Carin, Lawrence. IEEE Transactions on Image Processing . 2011

[8]

Subspace learning for background modeling: A survey. Bouwmans,Thierry. Recent Patents on Computer Science . 2009

[9]

Robust Recovery of Subspace Structures by Low-Rank Representation. Guangcan Liu,Zhouchen Lin,Shuicheng Yan,Ju Sun,Yong Yu,Yi Ma. IEEE Transactions on Pattern Analysis and Machine Intelligence . 2013

[10]

The Augmented Lagrange Multiplier Methodfor Exact Recovery of Corrupted Low-rank Matrices. Lin Zhouchen,Chen Minming,Wu Leqin,Ma Yi. UIUC Technical ReportUILU-ENG-09-2212 . 2009

← 1 2 →