ICASSP 2016

Neural network based spectral mask estimation for acoustic beamforming

ICASSP 2016 fgnt/nn-gev

The network training is independent of the number and the geometric configuration of the microphones.

SPEECH RECOGNITION