DNN-based speech mask estimation for eigenvector beamforming.
Lukas PfeifenbergerMatthias ZöhrerFranz PernkopfPublished in: ICASSP (2017)
Keyphrases
- speech enhancement
- speech recognition
- robust estimation
- blind source separation
- pattern recognition
- linear combination
- speech synthesis
- accurate estimation
- audio visual
- mimo ofdm systems
- endpoint detection
- doa estimation
- vocal tract
- spoken language
- data sets
- estimation algorithm
- estimation error
- density estimation
- principal component analysis
- digital images
- neural network