Increasing virtual samples through loss smoothness determination in large geometric margin minimum classification error training.
Tsukasa OhashiHideyuki WatanabeJun'ichi TokunoShigeru KatagiriMiho OhsakiShigeki MatsudaHideki KashiokaPublished in: ICASSP (2012)
Keyphrases
- minimum classification error
- training set
- statistical pattern recognition
- training samples
- discriminative training
- classification error
- maximum likelihood estimation
- objective function
- hidden markov models
- data sets
- training examples
- gaussian mixture
- pattern recognition
- support vector
- active learning
- class labels
- feature extraction
- decision trees
- image processing