Fig. 5From: Speech dereverberation for enhancement and recognition using dynamic features constrained deep neural networks and feature adaptationIllustration of cross transformBack to article page