Abstract:Recently, L 1 -norm-based robust discriminant feature extraction technique has been attracted much attention in dimensionality reduction and pattern recognition. However, it does not relate to the scatter matrix which well characterizes the geometric structure of data. In this paper, we propose a robust formulation of graph embedding framework for dimensionality reduction. In this robust framework, we use L 2 -norm to measure the distance along space dimension and L 1 -norm to sum overall data points. The proposed robust graph embedding framework retains the traditional framework's desirable properties, such as rotational invariance and well geometric structure, and simultaneously suppresses outliers. Based on this framework, we develop a simple and robust feature extraction method, namely L 2,1 -norm-based discriminant locality preserving projections (L 2,1 -DLPP) and provide an effective iterative algorithm to solve L 2,1 -DLPP. Extensive experiments in artificial data and three popular face databases illustrate the effectiveness of our proposed method.

A New Low-Rank Structurally Incoherent Algorithm for Robust Image Feature Extraction

Image Feature Optimization Based on Nonlinear Dimensionality Reduction

Low-rank 2D Local Discriminant Graph Embedding for Robust Image Feature Extraction

Robust non-negative supervised low-rank discriminant embedding (NSLRDE) for feature extraction

Structurally Incoherent Low-Rank 2DLPP for Image Classification

Structurally Incoherent Low-Rank Nonnegative Matrix Factorization for Image Classification

Learning Latent Low-Rank and Sparse Embedding for Robust Image Feature Extraction

Low-Rank Projection Learning Via Graph Embedding

Robust sparse low-rank embedding for image dimension reduction

Dual low-rank structure embedding for robust visual information processing

${L_{2,1}}$ -Norm Discriminant Manifold Learning

Local Neighborhood Embedding for Unsupervised Nonlinear Dimension Reduction.

Robust Integrated Locally Linear Embedding

Robust Image Annotation Via Simultaneous Feature and Sample Outlier Pursuit

Unsupervised Large Graph Embedding

Block-Diagonal Constrained Low-Rank and Sparse Graph for Discriminant Analysis of Image Data.

Discriminative feature extraction based on sparse and low-rank representation

Edge Structure Learning Via Low Rank Residuals for Robust Image Classification.

Robust Adaptive Low-Rank and Sparse Embedding for Feature Representation

<inline-formula> <tex-math notation="LaTeX">${L_{2,1}}$ </tex-math></inline-formula>-Norm Discriminant Manifold Learning

LGSLRR: Towards fusing discriminative ordinal local and global structured low-rank representation for image recognition