Online Multichannel Speech Enhancement Based on Recursive EM and DNN-Based Speech Presence Estimation.
Juan M. Martín-DoñasJesper JensenZheng-Hua TanAngel M. GomezAntonio M. PeinadoPublished in: IEEE ACM Trans. Audio Speech Lang. Process. (2020)
Keyphrases
- speech enhancement
- single channel
- linear prediction
- multi channel
- noisy environments
- speech signal
- noise reduction
- signal to noise ratio
- vocal tract
- noisy speech
- wiener filter
- independent component analysis
- background noise
- frequency domain
- prior information
- parameter estimation
- maximum likelihood
- speech recognition
- em algorithm
- automatic speech recognition
- lossless compression
- brain activity
- autoregressive
- speaker identification
- information retrieval
- prediction error
- expectation maximization
- probabilistic model