Constructing the L2-Graph for Robust Subspace Learning and Subspace Clustering
arXiv:1209.0841 · doi:10.1109/TCYB.2016.2536752
Abstract
Under the framework of graph-based learning, the key to robust subspace clustering and subspace learning is to obtain a good similarity graph that eliminates the effects of errors and retains only connections between the data points from the same subspace (i.e., intra-subspace data points). Recent works achieve good performance by modeling errors into their objective functions to remove the errors from the inputs. However, these approaches face the limitations that the structure of errors should be known prior and a complex convex problem must be solved. In this paper, we present a novel method to eliminate the effects of the errors from the projection space (representation) rather than from the input space. We first prove that -, -, -, and nuclear-norm based linear projection spaces share the property of Intra-subspace Projection Dominance (IPD), i.e., the coefficients over intra-subspace data points are larger than those over inter-subspace data points. Based on this property, we introduce a method to construct a sparse similarity graph, called L2-Graph. The subspace clustering and subspace learning algorithms are developed upon L2-Graph. Experiments show that L2-Graph algorithms outperform the state-of-the-art methods for feature extraction, image clustering, and motion segmentation in terms of accuracy, robustness, and time efficiency.
References in corpus (6)
- Robust Recovery of Subspace Structures by Low-Rank Representation
- Alternating Direction Algorithms for -Problems in Compressive Sensing
- Low-Rank Structure Learning via Log-Sum Heuristic Recovery
- Learning Locality-Constrained Collaborative Representation for Face Recognition
- Analysis and Improvement of Low Rank Representation for Subspace segmentation
- Noisy Sparse Subspace Clustering
Cited by in corpus (28)
- Robust Graph Learning from Noisy Data
- Structured Sparse Subspace Clustering: A Joint Affinity Learning and Subspace Clustering Framework
- Connections Between Nuclear Norm and Frobenius Norm Based Representations
- Pseudo-supervised Deep Subspace Clustering
- Zero-Shot Learning via Semantic Similarity Embedding
- Learning Locality-Constrained Collaborative Representation for Face Recognition
- -Motivated Low-Rank Sparse Subspace Clustering
- Subspace clustering using a symmetric low-rank representation
- Fast Kronecker product kernel methods via generalized vec trick
- A fast and Accurate Similarity-constrained Subspace Clustering Framework for Unsupervised Hyperspectral Image Classification
- Active Orthogonal Matching Pursuit for Sparse Subspace Clustering
- Symmetric low-rank representation for subspace clustering
- Deep Sparse Subspace Clustering
- Simultaneous Subspace Clustering and Cluster Number Estimating based on Triplet Relationship
- Discriminative Transformation Learning for Fuzzy Sparse Subspace Clustering
- Structured Graph Learning for Scalable Subspace Clustering: From Single-view to Multi-view
- Efficient Sparse Subspace Clustering by Nearest Neighbour Filtering
- Beyond Linear Subspace Clustering: A Comparative Study of Nonlinear Manifold Clustering Algorithms
- Revisiting data augmentation for subspace clustering
- Affinity Fusion Graph-based Framework for Natural Image Segmentation
- Scaled Simplex Representation for Subspace Clustering
- Label-independent hyperparameter-free self-supervised single-view deep subspace clustering
- Clustering-aware Graph Construction: A Joint Learning Perspective
- Robust Self-Supervised Convolutional Neural Network for Subspace Clustering and Classification
- Structure Learning with Similarity Preserving
- Kernel Truncated Regression Representation for Robust Subspace Clustering
- 3D Shape Classification Using Collaborative Representation based Projections
- Locally linear representation for image clustering