Emotional Video to Audio Transformation Using Deep Recurrent Neural Networks and a Neuro-Fuzzy System

Table 2

Extended MOS and MAE results with 8 samples.

	MAE	Target MOS		Obtained MOS
		Valence	Arousal	Valence	Arousal
1	0.193	8.0	8.6	4.43 ± 1.99	5.14 ± 2.10
2	0.142	3.8	5.2	5.57 ± 1.84	5.57 ± 1.59
3	0.234	6.4	8.0	4.71 ± 1.58	5.00 ± 1.78
4	0.203	6.4	6.8	4.71 ± 1.39	3.71 ± 1.03
5	0.280	6.6	7.0	6.43 ± 1.29	6.57 ± 1.05
6	0.262	6.0	6.8	3.00 ± 0.53	4.29 ± 0.88
7	0.206	7.2	6.8	5.86 ± 1.46	4.71 ± 2.05
8	0.219	5.2	7.2	5.43 ± 1.18	5.86 ± 1.36