Semi-tight covariance matrices implementation in MASPER HMM training procedure.

Juraj Kacur Robert Kozicka Radoslav Vargic

Published in: IWSSIP (2016)

Keyphrases

covariance matrices
training procedure
covariance matrix
maximum likelihood
distance measure
gaussian distribution
gaussian mixture model
vector space
neural network
gaussian mixture
principal component analysis
data sets
lower bound
support vector machine
decision trees
learning algorithm
multiple kernel learning