Hyperspectral image unmixing by non-negative matrix factorization initialized with modified independent component analysis.
Djaouad BenachirYannick DevilleShahram HosseiniMoussa Sofiane KarouiAbdelkader HameurlainPublished in: WHISPERS (2013)
Keyphrases
- negative matrix factorization
- independent component analysis
- hyperspectral images
- principal component analysis
- blind source separation
- subspace projection
- hyperspectral
- nonnegative matrix factorization
- hyperspectral imagery
- hyperspectral data
- factor analysis
- independent components
- remote sensing
- dimensionality reduction
- face recognition
- matrix factorization
- principal components
- document clustering
- sparse representation
- signal processing
- random projections
- feature space
- infrared
- low dimensional
- dimension reduction
- multispectral
- principle component analysis
- multispectral images
- spectral clustering
- least squares
- high resolution