A new two-stage method for single-microphone speech dereverberation.
Ali BaghakiM. Omair AhmadM. N. S. SwamyPublished in: ISCAS (2016)
Keyphrases
- experimental evaluation
- synthetic data
- cost function
- high accuracy
- multiscale
- computational complexity
- significant improvement
- bayesian networks
- high precision
- computational cost
- classification method
- speech synthesis
- feature selection
- text to speech
- fully automatic
- optimization algorithm
- support vector machine svm
- theoretical analysis
- probabilistic model
- dynamic programming
- prior knowledge
- pairwise
- similarity measure