Redesign of Gaussian Mixture Model for Efficient and Privacy-preserving Speaker Recognition.
S. RahulamathavanX. YaoR. YogachandranK. CumananMuttukrishnan RajarajanPublished in: CyberSA (2018)
Keyphrases
- gaussian mixture model
- speaker recognition
- privacy preserving
- mixture model
- privacy preservation
- privacy preserving data mining
- speaker identification
- maximum likelihood
- feature vectors
- vertically partitioned data
- em algorithm
- expectation maximization
- data privacy
- vector quantization
- feature space
- prior knowledge
- speaker verification
- secure multiparty computation
- probabilistic neural network
- data sets
- pattern recognition
- bayesian networks
- decision trees
- feature selection
- computer vision