EURASIP Journal on Audio, Speech, and Music Processing

Table 8 Recognition rates of DNN with varying numbers of units in each layer for training data (%)

From: Deep neural network-based bottleneck feature and denoising autoencoder-based dereverberation for distant-talking speaker identification

Method	Number of units	CENSREC-4 database			Ave.
	in each layer	0.40	0.60	0.75
BF-DNN	512	84.90	78.95	83.15	82.33
BF-DNN	1,024	88.40	83.90	88.10	86.80
BF-DNN	2048	87.45	83.25	87.05	85.92

Back to article page