中国计量大学学报2018，Vol.29Issue(4)：385-392,8.DOI:10.3969/j.issn.2096-2835.2018.04.007
基于加权Schatten-p范数与树结构稀疏分解的目标显著性检测
Salient object detection based on weighted Schatten-pnorms and tree structured sparsity decomposition
摘要
Abstract
Recently, salient object detection has attracted considerable attention;and some methods based on the low rank matrix recovery theory have been proposed.For these methods, people generally constrain low rank component with nuclear norms.However, since the rank function is non-convex and discontinuous, the nuclear norm may not be a good approximation.Therefore, the salient object detection does not achieve desirable results.To address this problem, a novel model based on weighted Schatten-p norms and tree structured sparsity decomposition was proposed.The weighted Schatten-p norm was used to constrain the background of the image.On the other hand, the salient object was constrained by the tree structured sparsity l2, 1norm and Laplacian regularization to improve the accuracy of saliency detection.Extensive experimental results showed that the proposed method has a better detection performance compared with four state-of-theart detection methods on three different databases.关键词
目标显著性检测/矩阵分解/加权Schatten-p范数/树结构/拉普拉斯正则化Key words
salient object detection/matrix decomposition/weighted Schatten-p norm/tree structure/Laplacian regularization分类
信息技术与安全科学引用本文复制引用
钱文超,曹飞龙..基于加权Schatten-p范数与树结构稀疏分解的目标显著性检测[J].中国计量大学学报,2018,29(4):385-392,8.基金项目
国家自然科学基金项目(No.61672477) （No.61672477）
