HSR: L 1/2-regularized sparse representation for fast face recognition using hierarchical feature selection.
Bo HanBo HeTingting SunTianhong YanMengmeng MaYue ShenAmaury LendassePublished in: Neural Comput. Appl. (2016)
Keyphrases
- sparse representation
- feature selection
- regularized least squares
- sparse reconstruction
- dictionary learning
- dimensionality reduction
- norm minimization
- face recognition
- sparse coding
- sparsity constraints
- orthogonal matching pursuit
- face images
- negative matrix factorization
- image classification
- high dimensional data
- feature extraction
- high dimensionality
- image representation
- support vector
- facial expressions
- signal processing
- compressive sensing
- compressed sensing
- joint optimization
- unsupervised learning
- image patches
- random projections
- robust face recognition
- text classification
- least squares
- feature space
- subspace learning
- machine learning
- low dimensional
- sparse approximations