Frequency-dependent auto-pooling function for weakly supervised sound event detection

EURASIP Journal on Audio, Speech, and Music Processing

Table 7 The results of DCASE 2020 Task 4 dataset

	Parameters	Audio tagging			Sound event detection			Error rate
	10 k	F-score	AUC	mAP	F-score	AUC	mAP	ER	D	I
Attention [20]	54.15	0.719	0.946	0.810	0.557	0.867	0.572	1.715	0.755	0.960
TALNet [19]	94.06	0.672	0.917	0.741	0.536	0.851	0.516	1.523	0.773	0.750
VGG-GWRP [24]	58.76	0.675	0.938	0.772	0.499	0.843	0.578	1.866	0.769	1.096
DDC-FAP	29.10	0.694	0.941	0.795	0.595	0.881	0.610	1.755	0.790	0.965