EURASIP Journal on Audio, Speech, and Music Processing

Table 18 System results of the ALBAYZIN 2018 QbE STD evaluation on RTVE test data for single-word (Single) and multi-word (Multi) queries

From: Search on speech from spoken queries: the Multi-domain International ALBAYZIN 2018 Query-by-Example Spoken Term Detection Evaluation

	Single				Multi
System ID	MTWV	ATWV	p(FA)	p(Miss)	MTWV	ATWV	p(FA)	p(Miss)
A-Hybrid DTW+LVCSR	0.6122	0.5651	0.00008	0.310	0.9531	0.7374	0.00005	0.000
B-Fusion DTW	0.4369	0.3849	0.00004	0.523	0.4656	0.4093	0.00000	0.534
C-PhonePost DTW	0.4167	0.3374	0.00002	0.565	0.4562	0.4281	0.00001	0.534
D-LVCSR	0.4565	0.3854	0.00008	0.460	0.9906	0.8218	0.00001	0.000
E-DTW	0.0000	−0.0145	0.00000	1.000	0.0000	0.0000	0.00000	1.000
F-Combined DTW	0.3066	0.2904	0.00003	0.664	0.8860	0.8485	0.00002	0.095
G-Super-BNF DTW	0.2879	0.2849	0.00004	0.670	0.8672	0.8290	0.00004	0.095
H-Multilingual-BNF DTW	0.2765	0.2724	0.00005	0.677	0.8478	0.7915	0.00001	0.143
I-Monoph.-BNF DTW	0.3073	0.3070	0.00005	0.642	0.9055	0.9055	0.00005	0.048
J-Triph.-BNF DTW	0.2960	0.2808	0.00004	0.668	0.8583	0.8302	0.00001	0.132
K-Text STD	0.6917	0.6625	0.00008	0.229	0.8783	0.8783	0.00000	0.122

Back to article page