Speech Enhancement on LRS3 mixed with QUT city-street noises (test) benchmark leaderboard · SOTA2 Research

Speech Enhancement on LRS3 mixed with QUT city-street noises (test)

3.21PESQVisual-stream method [2]

Evaluation Results

Method
Visual-stream method [2]2020.12	3.21	3.48	2.77	2.45	0.95
pseudo-visual speech enhancement2020.12	3.12	3.42	2.73	2.41	0.94
AO (Audio-Only)2020.12	3.05	3.35	2.62	2.35	0.91
Visual-stream method [2]2020.12	3.01	3.33	2.63	2.34	0.93
pseudo-visual speech enhancement2020.12	2.92	3.24	2.59	2.29	0.91
Method [11]2020.12	2.9	3.38	2.59	2.3	0.91
AO (Audio-Only)2020.12	2.83	3.15	2.49	2.2	0.89
Visual-stream method [2]2020.12	2.73	3.1	2.46	2.2	0.89
Noisy2020.12	2.67	3.3	2.42	2.14	0.87
pseudo-visual speech enhancement2020.12	2.65	3.02	2.43	2.12	0.87
Method [19]2020.12	2.61	3.3	2.53	2.2	0.9
AO (Audio-Only)2020.12	2.54	2.93	2.31	2.04	0.84
Method [11]2020.12	2.54	3.18	2.3	2.06	0.87
Noisy2020.12	2.26	2.9	1.94	2.02	0.85
Method [11]2020.12	2.14	2.78	2.1	1.95	0.8
Method [19]2020.12	2.14	2.86	2.23	1.99	0.86
Noisy2020.12	1.86	2.46	1.55	1.69	0.75
Method [19]2020.12	1.77	2.37	1.89	1.68	0.77