《计算机应用》唯一官方网站 ›› 2024, Vol. 44 ›› Issue (2): 353-359.DOI: 10.11772/j.issn.1001-9081.2023030275

• 人工智能 • 上一篇    

基于一致性和多样性的多尺度自表示学习的深度子空间聚类

张卓, 陈花竹()   

  1. 中原工学院 理学院,郑州 450007
  • 收稿日期:2023-03-15 修回日期:2023-05-26 接受日期:2023-05-30 发布日期:2023-06-30 出版日期:2024-02-10
  • 通讯作者: 陈花竹
  • 作者简介:张卓(1998—),男,河南商丘人,硕士研究生 ,主要研究方向:图像处理;
  • 基金资助:
    国家自然科学基金资助项目(62202513);中原工学院基本科研业务费专项资金资助项目(K2022YY012)

Deep subspace clustering based on multiscale self-representation learning with consistency and diversity

Zhuo ZHANG, Huazhu CHEN()   

  1. College of Science,Zhongyuan University of Technology,Zhengzhou Henan 450007,China
  • Received:2023-03-15 Revised:2023-05-26 Accepted:2023-05-30 Online:2023-06-30 Published:2024-02-10
  • Contact: Huazhu CHEN
  • About author:ZHANG Zhuo, born in 1998, M. S. candidate. His research interests include image processing.
  • Supported by:
    National Natural Science Foundation of China(62202513);Basic Research Fund for Zhongyuan University of Technology(K2022YY012)

摘要:

深度子空间聚类(DSC)基于原始数据位于低维非线性子空间的集合中的假设。其中深度子空间聚类多尺度表示学习方法在深度自编码器的基础上,将每一层的编码器与对应的解码器之间都添加全连接层,并以此捕获多尺度的特征,但它没有深度分析多尺度特征的性质,也没有考虑输入数据和输出数据之间多尺度的重构损失。为了解决上述问题,首先建立每个网络层的重构损失函数,监督不同级别编码器参数的学习;然后利用多尺度特征共有的自表示矩阵和特有的自表示矩阵的和具有块对角性,提出更有效的多尺度自表示模块;最后分析不同尺度特征特有的自表示矩阵之间的多样性,有效地利用了多尺度的特征矩阵。在此基础上,提出一种基于一致性和多样性的多尺度自表示学习的深度子空间聚类(MSCD-DSC)方法。在数据集Extended Yale B 、ORL、COIL20和Umist上的实验结果表明,相较于次优的MLRDSC(Multi-Level Representation learning for Deep Subspace Clustering),MSCD-DSC的聚类错误率分别降低了15.44%、2.22%、3.37%和13.17%,表明MSCD-DSC的聚类效果优于已有的方法。

关键词: 深度子空间聚类, 自编码器, 多尺度, 自表示矩阵, 一致性, 多样性

Abstract:

Deep Subspace Clustering (DSC) is based on the assumption that the original data lies in a collection of low-dimensional nonlinear subspaces. In the multi-scale representation learning methods for deep subspace clustering, based on deep auto-encoder, fully connected layers are added between the encoder and the corresponding decoder for each layer to capture multi-scale features, without deeply analyzing the nature of multi-scale features and considering the multi-scale reconstruction loss between input data and output data. In order to solve the above problems, firstly, the reconstruction loss function of each network layer was established to supervise the learning of encoder parameters at different levels; then, a more effective multi-scale self-representation module was proposed based on the block diagonality of the sum of the common self-representation matrix and the unique self-representation matrices for multi-scale features; finally, the diversity of unique self-representation matrices for different scale features was analyzed in depth and the multi-scale feature matrices were used effectively. On this basis, an MSCD-DSC (Multiscale Self-representation learning with Consistency and Diversity for Deep Subspace Clustering) method was proposed. Experimental results on the datasets Extended Yale B, ORL, COIL20 and Umist show that, compared to the suboptimal method MLRDSC (Multi-Level Representation learning for Deep Subspace Clustering), the clustering error rate of MSCD-DSC is reduced by 15.44%, 2.22%, 3.37%, and 13.17%, respectively, indicating that the clustering effect of MSCD-DSC is better than those of the existing methods.

Key words: deep subspace clustering, auto-encoder, multiscale, self-representation matrix, consistency, diversity

中图分类号: