Comparison of input and feature space nonlinear kernel nuisance attribute projections for speaker verification.
Xianyu ZhaoYuan DongJian ZhaoLiang LuJiqing LiuHaila WangPublished in: INTERSPEECH (2008)
Keyphrases
- speaker verification
- feature space
- nonlinear kernel
- noisy environments
- input data
- canonical correlation analysis
- high dimensional
- emotion recognition
- classification accuracy
- kernel function
- input space
- feature vectors
- data points
- principal component analysis
- dimensionality reduction
- metric learning
- feature extraction
- feature selection
- face verification
- data sets
- training data
- svm classification
- computer vision