Zhou Nan, Choi Kup-Sze, Chen Badong, Du Yuanhua, Liu Jun, Xu Yangyang
IEEE Trans Neural Netw Learn Syst. 2023 Dec;34(12):10433-10446. doi: 10.1109/TNNLS.2022.3166931. Epub 2023 Nov 30.
This article proposes a novel low-rank matrix factorization model for semisupervised image clustering. In order to alleviate the negative effect of outliers, the maximum correntropy criterion (MCC) is incorporated as a metric to build the model. To utilize the label information to improve the clustering results, a constraint graph learning framework is proposed to adaptively learn the local structure of the data by considering the label information. Furthermore, an iterative algorithm based on Fenchel conjugate (FC) and block coordinate update (BCU) is proposed to solve the model. The convergence properties of the proposed algorithm are analyzed, which shows that the algorithm exhibits both objective sequential convergence and iterate sequential convergence. Experiments are conducted on six real-world image datasets, and the proposed algorithm is compared with eight state-of-the-art methods. The results show that the proposed method can achieve better performance in most situations in terms of clustering accuracy and mutual information.
本文提出了一种用于半监督图像聚类的新型低秩矩阵分解模型。为了减轻离群值的负面影响，引入最大相关熵准则（MCC）作为一种度量来构建模型。为了利用标签信息来改善聚类结果，提出了一种约束图学习框架，通过考虑标签信息来自适应地学习数据的局部结构。此外，还提出了一种基于Fenchel共轭（FC）和块坐标更新（BCU）的迭代算法来求解该模型。分析了所提算法的收敛性质，结果表明该算法具有目标序列收敛性和迭代序列收敛性。在六个真实世界图像数据集上进行了实验，并将所提算法与八种最新方法进行了比较。结果表明，所提方法在大多数情况下，在聚类精度和互信息方面都能取得更好的性能。
