EURASIP Journal on Audio, Speech, and Music Processing

Table 10 Recognition rates of BF-DNN with a varying number of layers for training data (%)

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

Method	Number of layers	CENSREC-4 database			Ave.
		0.40	0.60	0.75
BF-DNN	3	56.60	48.65	55.45	53.57
BF-DNN	5	88.40	83.90	88.10	86.80
BF-DNN	7	90.40	86.40	89.20	88.67
BF-DNN	9	91.80	88.35	91.15	90.43

Back to article page