Speech Enhancement on LRS3 mixed with VGGSound noises (test) benchmark leaderboard · SOTA2 Research

Speech Enhancement on LRS3 mixed with VGGSound noises (test)

3.25PESQVisual-stream method [2]

Evaluation Results

Method
Visual-stream method [2]2020.12	3.25	3.56	2.84	2.58	0.95
pseudo-visual speech enhancement2020.12	3.19	3.51	2.81	2.52	0.95
AO (Audio-Only)2020.12	3.12	3.45	2.7	2.46	0.92
Visual-stream method [2]2020.12	3.05	3.39	2.71	2.41	0.94
pseudo-visual speech enhancement2020.12	2.99	3.32	2.65	2.37	0.92
Method [11]2020.12	2.95	3.36	2.63	2.31	0.9
AO (Audio-Only)2020.12	2.93	3.26	2.54	2.29	0.9
Visual-stream method [2]2020.12	2.8	3.25	2.51	2.29	0.9
pseudo-visual speech enhancement2020.12	2.72	3.18	2.47	2.25	0.88
Noisy2020.12	2.66	3.15	2.4	2.16	0.88
Method [19]2020.12	2.65	3.17	2.55	2.26	0.88
AO (Audio-Only)2020.12	2.62	3.02	2.36	2.14	0.87
Method [11]2020.12	2.52	3.22	2.46	2.21	0.88
Noisy2020.12	2.29	2.79	2.23	2.04	0.85
Method [19]2020.12	2.24	2.67	2.3	2	0.86
Method [11]2020.12	2.17	2.82	2.3	2.02	0.83
Noisy2020.12	1.93	2.31	1.83	1.68	0.76
Method [19]2020.12	1.84	2.1	1.87	1.53	0.75