Extension of the remos concept to frequency-filtering-based features for reverberation-robust Speech recognition | IEEE Conference Publication | IEEE Xplore