A frequency-selecting multichannel equalization method for speech dereverberation.
Wancheng ZhangYiming YanTianyi WangZunwen HePublished in: DSP (2015)
Keyphrases
- high accuracy
- cost function
- experimental evaluation
- similarity measure
- high precision
- error rate
- detection method
- prior knowledge
- computational cost
- probabilistic model
- preprocessing
- synthetic data
- significant improvement
- pairwise
- edge detection
- support vector machine svm
- detection algorithm
- clustering method
- support vector
- segmentation method
- genetic algorithm