Combining Multiple Kernel Methods on Riemannian Manifold for Emotion Recognition in the Wild.
Mengyi LiuRuiping WangShaoxin LiShiguang ShanZhiwu HuangXilin ChenPublished in: ICMI (2014)
Keyphrases
- combining multiple
- kernel methods
- riemannian manifolds
- reproducing kernel hilbert space
- feature space
- euclidean space
- kernel function
- kernel pca
- kernel matrix
- learning problems
- support vector
- human computer interaction
- facial images
- facial expressions
- geometric structure
- machine learning
- learning tasks
- vector space
- information fusion
- support vector machine
- graph kernels
- metric space
- covariance matrix
- input data
- image classification
- face images
- geodesic distance
- classification accuracy
- feature vectors
- pairwise
- video sequences
- data sets