Nuclear Norm Based Matrix Regression with Applications to Face Recognition with Occlusion and Illumination Changes.
Jian YangLei LuoJianjun QianYing TaiFanlong ZhangYong XuPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2017)
Keyphrases
- nuclear norm
- face recognition
- pose variations
- low rank
- illumination variations
- low rank matrix
- singular values
- norm minimization
- face images
- matrix completion
- low rank matrices
- sparse representation
- convex optimization
- minimization problems
- singular value decomposition
- rank minimization
- matrix factorization
- linear combination
- missing data
- high dimensional data
- partial occlusion
- feature extraction
- principal component analysis
- semi supervised
- sparse matrix
- lighting conditions
- facial expressions
- computer vision
- high order
- data matrix
- kernel matrix
- support vector
- approximation methods
- linear subspace
- high dimensional
- convex relaxation
- affinity matrix
- dimensionality reduction
- total variation