A DNN beamformer trained with an augmented Lagrangian-inspired adaptive loss using target RTF and interference subspace estimates outperforms classical LCMV beamformers in multi-speaker enhancement with better sidelobe control and noise attenuation.
A consolidated perspective on multimicrophone speech enhance- ment and source separation,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.AS 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Linearly Constrained Deep Beamformer for Multi-Speaker Scenarios
A DNN beamformer trained with an augmented Lagrangian-inspired adaptive loss using target RTF and interference subspace estimates outperforms classical LCMV beamformers in multi-speaker enhancement with better sidelobe control and noise attenuation.