Abstract:This paper proposes a novel nonnegative sparse representation approach, called two-stage sparse representation (TSR), for robust face recognition on a large-scale database. Based on the divide and conquer strategy, TSR decomposes the procedure of robust face recognition into outlier detection stage and recognition stage. In the first stage, we propose a general multisubspace framework to learn a robust metric in which noise and outliers in image pixels are detected. Potential loss functions, including L1 , L2,1, and correntropy are studied. In the second stage, based on the learned metric and collaborative representation, we propose an efficient nonnegative sparse representation algorithm to find an approximation solution of sparse representation. According to the L1 ball theory in sparse representation, the approximated solution is unique and can be optimized efficiently. Then a filtering strategy is developed to avoid the computation of the sparse representation on the whole large-scale dataset. Moreover, theoretical analysis also gives the necessary condition for nonnegative least squares technique to find a sparse solution. Extensive experiments on several public databases have demonstrated that the proposed TSR approach, in general, achieves better classification accuracy than the state-of-the-art sparse representation methods. More importantly, a significant reduction of computational costs is reached in comparison with sparse representation classifier; this enables the TSR to be more suitable for robust face recognition on a large-scale dataset.

Beyond sparsity: The role of L1-optimizer in pattern classification

Two-stage nonnegative sparse representation for large-scale face recognition.

Collaborative Representation for Classification, Sparse or Non-sparse?

Robust Face Recognition via Adaptive Sparse Representation

Robust and Practical Face Recognition Via Structured Sparsity

Fast Low Rank Representation Based Spatial Pyramid Matching for Image Classification.

Collaborative Representation based Classification for Face Recognition

Multiple Kernel Sparse Representation Based Classification

An Overview of Computational Sparse Models and Their Applications in Artificial Intelligence

K Nearest Neighbor Based Local Sparse Representation Classifier

Online Kernel Learning with a Near Optimal Sparsity Bound

Sparse Group Regression Classification for Face Recognition

Registration Method of Sparse Representation Classification Method.

Sparse non-negative transition subspace learning for image classification

Neighborhood Preserved Sparse Representation for Robust Classification on Symmetric Positive Definite Matrices

Adaptive Optimizers with Sparse Group Lasso for Neural Networks in CTR Prediction

Superimposed Sparse Parameter Classifiers for Face Recognition

Data Classification Algorithm Based on Sparse Decomposition

Dictionary Construction for Sparse Representation Classification: A Novel Cluster-Based Approach

Low-Rank Sparse Coding for Image Classification

A novel kernel collaborative representation approach for image classification