Speech Privacy for Sound Surveillance Using Super-Resolution Based on Maximum Likelihood and Bayesian Linear Regression.
Ryouichi NishimuraSeigo EnomotoHiroaki KatoPublished in: IEICE Trans. Inf. Syst. (2018)
Keyphrases
- linear regression
- super resolution
- maximum likelihood
- low resolution
- least squares
- image reconstruction
- high resolution
- image restoration
- em algorithm
- low resolution images
- speech recognition
- image super resolution
- regression methods
- motion estimation
- high quality
- high resolution images
- nonlinear regression
- super resolved
- expectation maximization
- super resolution reconstruction
- depth map
- maximum a posteriori
- speech signal
- multi frame
- higher resolution
- resolution enhancement
- super resolution algorithm
- kernel regression
- gaussian mixture model
- hyperparameters
- feature space
- video sequences