Skip to main content
EURASIP Journal on Audio, Speech, and Music Processing

Table 1 The FLOPs per frame of CRN, DCN, GRN, DARCN, and the 3 proposed posterfiltering methods

From: Low-complexity artificial noise suppression methods for deep learning-based speech enhancement algorithms

  DNNs Postfiltering methods
Method CRN DCN GRN DARCN SPP-proposed-1 SPP-proposed-2 SPP-proposed-3
MFLOPs 32.22 46.68 10.40 64.66 0.014 0.0098 0.016
Back to article page