Two-layer similarity fusion model for cover song identification

Chen, Ning; Li, Mingyu; Xiao, Haidong

doi:10.1186/s13636-017-0108-2

EURASIP Journal on Audio, Speech, and Music Processing

Table 5 Identification accuracy achieved by different descriptor combinations

From: Two-layer similarity fusion model for cover song identification

	Combination of HPCP and MLD					Combination of BSC and CPCP
		MAP	MaRR	TOP10	MR		MAP	MaRR	TOP10	MR
DB801	2L-Best1	0.6064	0.4748	1394	52	2L-Best1	0.5563	0.4396	1251	72
	2L-HPCP-Best1	0.5898	0.4628	1249	61	2L-BSC-Best1	0.5394	0.4255	1233	88
	1L-HPCP-QD [32]	0.5941	0.2549	1329	61	1L-BSC-QD	0.5415	0.2344	1207	78
	1L-MLD-QD	0.2825	0.1287	663	136	1L-CPCP-QD	0.1776	0.0887	403	174
	[26]	0.5393	0.2358	1138	84	[26]	0.5393	0.2358	1138	84
	[24]	–	–	–	–	[24]	–	–	–	–
	PSO	0.4899	0.2201	1030	77	PSO	0.4732	0.2115	1031	86
	[31]	0.3939	0.1842	859	77	[31]	0.3242	0.1553	763	90
DB799	2L-Best1	0.6105	0.4901	1326	50	2L-Best1	0.5280	0.4419	1105	78
	2L-HPCP-Best1	0.5824	0.4676	1268	57	2L-BSC-Best1	0.5050	0.4204	1049	90
	1L-HPCP-QD [32]	0.5873	0.2573	1266	56	1L-BSC-QD	0.5080	0.2286	1041	83
	1L-MLD-QD	0.2922	0.1362	605	140	1L-CPCP-QD	0.1693	0.086	324	189
	[26]	0.5418	0.2313	1122	78	[26]	0.5418	0.2313	1122	78
	[24]	–	–	–	–	[24]	–	–	–	–
	PSO	0.5025	0.2280	1058	70	PSO	0.4491	0.2165	8671	99
	[31]	0.4281	0.2009	894	67	[31]	0.2963	0.1460	644	96
DB802	2L-Best1	0.5764	0.4668	1233	52	2L-Best1	0.5461	0.4469	1169	67
	2L-HPCP-Best1	0.5609	0.4515	1208	66	2L-BSC-Best1	0.5166	0.4210	1143	96
	1L-HPCP-QD [32]	0.5635	0.2469	1201	55	1L-BSC-QD	0.5234	0.2327	1140	83
	1L-MLD-QD	0.2856	0.1337	636	134	1L-CPCP-QD	0.1757	0.0870	361	192
	[26]	0.5228	0.2318	1072	82	[26]	0.5228	0.2318	1072	82
	[24]	–	–	–	–	[24]	–	–	–	–
	PSO	0.5050	0.2277	1026	63	PSO	0.4401	0.2029	920	85
	[31]	0.4280	0.1986	896	61	[31]	0.3196	0.1552	728	86
DB962	2L-Best1	0.6315	0.5440	1347	87	2L-Best1	0.5833	0.5186	1124	105
	2L-HPCP-Best1	0.6125	0.5242	1334	102	2L-BSC-Best1	0.5567	0.4921	1098	129
	1L-HPCP-QD [32]	0.6135	0.2801	1316	92	1L-BSC-QD	0.5573	0.2602	1073	105
	1L-MLD-QD	0.3417	0.1683	662	174	1L-CPCP-QD	0.2238	0.1134	389	236
	[26]	0.5856	0.2710	1170	105	[26]	0.5856	0.2710	1170	105
	[24]	–	–	–	–	[24]	–	–	–	–
	PSO	0.5452	0.2568	1098	91	PSO	0.4851	0.2336	908	111
	[31]	0.4568	0.2213	917	89	[31]	0.3627	0.1819	699	113

Notes: For MAP, MaRR, and TOP10, the larger the value is, the better performance is obtained. For MR, the smaller the value is, the better performance is obtained

Back to article page