Spectral compression: Weighted principal component analysis versus weighted least squares.
Farnaz AgahianBrian V. FuntSeyed Hossein AmirshahiPublished in: Human Vision and Electronic Imaging (2014)
Keyphrases
- weighted least squares
- principal component analysis
- least squares
- missing data
- conjugate gradient
- error function
- motion parameters
- dimensionality reduction
- face recognition
- generalization error
- low dimensional
- face images
- data sets
- training data
- levenberg marquardt
- feature extraction
- deterministic annealing
- image sequences