Text-Prompted Multistep Speaker Verification Using Gibbs-Distribution-Based Extended Bayesian Inference for Reducing Verification Errors.
Shuichi KurogiTakuya UekiYuta MizobeTakeshi NishidaPublished in: ICONIP (3) (2013)
Keyphrases
- bayesian inference
- speaker verification
- probabilistic model
- face verification
- prior information
- noisy environments
- hyperparameters
- language identification
- audio visual
- text mining
- emotion recognition
- computer vision
- hidden variables
- multilayer perceptron
- particle filter
- image processing
- neural network
- multi modal
- denoising
- input image
- multiscale
- face recognition
- gibbs distribution