Deep Structure and Attention Aware Subspace Clustering
This addresses clustering tasks in computer vision and pattern recognition, offering an incremental improvement by incorporating structure information into deep clustering.
The paper tackles the problem of deep clustering by proposing a method that simultaneously considers data content and structure information, using a vision transformer to extract features and learning a subspace structure for spectral clustering, and demonstrates significant outperformance over state-of-the-art methods in experiments.
Clustering is a fundamental unsupervised representation learning task with wide application in computer vision and pattern recognition. Deep clustering utilizes deep neural networks to learn latent representation, which is suitable for clustering. However, previous deep clustering methods, especially image clustering, focus on the features of the data itself and ignore the relationship between the data, which is crucial for clustering. In this paper, we propose a novel Deep Structure and Attention aware Subspace Clustering (DSASC), which simultaneously considers data content and structure information. We use a vision transformer to extract features, and the extracted features are divided into two parts, structure features, and content features. The two features are used to learn a more efficient subspace structure for spectral clustering. Extensive experimental results demonstrate that our method significantly outperforms state-of-the-art methods. Our code will be available at https://github.com/cs-whh/DSASC