《计算机应用研究》|Application Research of Computers

基于低秩矩阵二元分解的快速显著性目标检测算法

Efficient salient object detection via low-rank matrix bi-factorization

免费全文下载 (已被下载 次)  
获取PDF全文
作者 刘明明,仇文宁,孙伟
机构 1.江苏建筑职业技术学院 智能制造学院,江苏 徐州 221008;2.中国矿业大学 信控学院,江苏 徐州 221116
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2020)07-061-2210-07
DOI 10.19734/j.issn.1001-3695.2018.11.0911
摘要 近年来,基于矩阵低秩表示模型的图像显著性目标检测受到了广泛关注。在传统模型中通常对秩最小化问题进行凸松弛,但是这种方法在每次迭代中必须执行矩阵奇异值分解(SVD),计算复杂度较高。为此,提出了一种低秩矩阵双因子分解和结构化稀疏矩阵分解联合优化模型,并应用于显著性目标检测。该模型不仅利用低秩矩阵双因子分解和交替方向法(ADM)来降低时间开销,而且引入分层稀疏正则化刻画稀疏矩阵中元素之间的空间关系;此外,所提算法能够无缝集成高层先验知识指导矩阵分解过程。实验结果表明,提出的算法检测性能优于当前主流无监督显著性目标检测算法,且具有较低的时间复杂度。
关键词 显著性目标检测; 低秩矩阵双因子分解; 分层稀疏正则化; 交替方向法
基金项目 国家自然科学基金资助项目(61801198)
江苏省自然科学基金资助项目(BK20180174)
“青蓝工程”资助项目
本文URL http://www.arocmag.com/article/01-2020-07-061.html
英文标题 Efficient salient object detection via low-rank matrix bi-factorization
作者英文名 Liu Mingming, Qiu Wenning, Sun Wei
机构英文名 1.School of Intelligent Manufacturing,Jiangsu Vocational Institute of Architectural Technology,Xuzhou Jiangsu 221008,China;2.School of Information & Control Engineering,China University of Mining & Technology,Xuzhou Jiangsu 221116,China
英文摘要 In recent years, salient object detection via low-rank recovery models has received a significant amount of attention in the field of object detection. Traditional models generally decompose an original image into a low-rank matrix and a sparse matrix by minimizing the nuclear norm. But these methods suffer from high computation complexity due to singular value decomposition(SVD). To solve this issue, this paper presented an efficient low-rank matrix bi-factorization model for salient object detection, which not only took advantage of low-rank matrix bi-factorization and alternating direction method(ADM) to reduce the computation cost, but utilized structured-sparsity regularization to exploit the spatial relations between the elements in the sparse matrix. Furthermore, this paper introduced high-level priors to jointly guide the matrix decomposition. Experimental results on five challenging datasets validate the proposed method outperforms the state-of-the-art methods in terms of six performance metrics.
英文关键词 salient object detection; low-rank matrix bi-factorization; hierarchical sparse regularization; alternating direction method
参考文献 查看稿件参考文献
 
收稿日期 2018/11/17
修回日期 2019/1/3
页码 2210-2216
中图分类号 TP391.41
文献标志码 A