Abstract:Sparse representation is a mathematical model for data representation that has proved to be a powerful tool for solving problems in various fields such as pattern recognition, machine learning, and computer vision. As one of the building blocks of the sparse representation method, dictionary learning plays an important role in the minimization of the reconstruction error between the original signal and its sparse representation in the space of the learned dictionary. Although using training samples directly as dictionary bases can achieve good performance, the main drawback of this method is that it may result in a very large and inefficient dictionary due to noisy training instances. To obtain a smaller and more representative dictionary, in this paper, we propose an approach called Laplacian sparse dictionary (LSD) learning. Our method is based on manifold learning and double sparsity. We incorporate the Laplacian weighted graph in the sparse representation model and impose the l1-norm sparsity on the dictionary. An LSD is a sparse overcomplete dictionary that can preserve the intrinsic structure of the data and learn a smaller dictionary for each class. The learned LSD can be easily integrated into a classification framework based on sparse representation. We compare the proposed method with other methods using three benchmark-controlled face image databases, Extended Yale B, ORL, and AR, and one uncontrolled person image dataset, i-LIDS-MA. Results show the advantages of the proposed LSD algorithm over state-of-the-art sparse representation based classification methods.

Noise-robust Semi-Supervised Learning Via Fast Sparse Coding

Noise-Robust Semi-Supervised Learning by Large-Scale Sparse Coding.

Robust Image Analysis by L1-Norm Semi-supervised Learning

Semi-supervised Image Classification Via Nonnegative Least-Squares Regression

A Label Noise Robust Stacked Auto-Encoder Algorithm for Inaccurate Supervised Classification Problems

Graph Regularized Sparse Coding for Image Representation

Semi-Supervised Sparse Coding

Learnable Subspace Orthogonal Projection for Semi-supervised Image Classification

Semi-supervised Discriminant Analysis Based on Sparse-Coding Theory

Robust Graph-Based Semisupervised Learning for Noisy Labeled Data Via Maximum Correntropy Criterion

Towards Robust Graph Neural Networks against Label Noise

Laplacian Sparse Dictionary Learning for Image Classification Based on Sparse Representation

Variance-enlarged Poisson Learning for Graph-based Semi-Supervised Learning with Extremely Sparse Labeled Data

A locally weighted sparse graph regularized Non-Negative Matrix Factorization method.

A Noise-resistant Graph Neural Network by Semi-supervised Contrastive Learning

Probabilistic Semi-Supervised Learning via Sparse Graph Structure Learning

Noise-robust classification with hypergraph neural network

Robust Graph Learning From Noisy Data

Learning with Noisy Labels Via Sparse Regularization

Unsupervised feature extraction by low-rank and sparsity preserving embedding

LaplaceConfidence: A graph-based approach for learning with noisy labels