Impact of sparse representation on the admissible solutions of spectral unmixing by non-negative matrix factorization.
Neeraj KumarSaïd MoussaouiJérôme IdierDavid BriePublished in: WHISPERS (2015)
Keyphrases
- sparse representation
- negative matrix factorization
- nonnegative matrix factorization
- blind source separation
- dictionary learning
- sparse coding
- image classification
- face recognition
- hyperspectral
- reconstruction error
- random projections
- dimensionality reduction
- image representation
- high dimensional data
- regularized least squares
- image patches
- basis vectors
- subspace learning
- sparsity constraints
- matrix factorization
- neural network
- linear combination
- co occurrence
- image retrieval
- multiscale
- sparse coefficients