• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Chinese conference on pattern recognition and computer vision >Locally Consistent Constrained Concept Factorization with L_p Smoothness for Image Representation
【24h】

Locally Consistent Constrained Concept Factorization with L_p Smoothness for Image Representation

机译：与图像表示的L_P平滑度局部一致的约束概念分解

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Matrix factorization based on image representation algorithms have been widely used to deal with high-dimensional data. Previous studies have shown that matrix factorization methods can achieve remarkable performances in clustering. In this paper, we propose a novel method for image representation, called Locally Consistent Constrained Concept Factorization with L_p Smoothness (LCCCF-LS). The main contributions of our proposed LCCCF-LS method mainly include as follows: Firstly, the local geometric structure of the data is effectively explored using a graph regularizer. Secondly, the label information of the concepts is consistent with known label information without additional parameters. Finally, we add the L_p smoothness constraint to produce a smooth and more accurate solution, and thus ensure the smoothness of the coefficient matrix. Comprehensive experiments on several image datasets manifest the superiority of the proposed LCCCF-LS method.
机译：基于图像表示算法的矩阵分解已被广泛用于处理高维数据。以前的研究表明，矩阵分解方法可以实现聚类的显着性能。在本文中，我们提出了一种用于图像表示的新方法，称为L_P平滑度（LCCCF-LS）的局部一致的受约束概念分解。我们提出的LCCCF-LS方法的主要贡献主要包括如下：首先，使用图形规范器有效地探索数据的本地几何结构。其次，概念的标签信息与已知标签信息一致，没有其他参数。最后，我们添加了L_P平滑度约束来产生平滑且更准确的解决方案，从而确保系数矩阵的平滑度。综合实验在几个图像数据集中表现出所提出的LCCCF-LS方法的优越性。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号