Table 3 Average standard MOS results with 95 % confidence intervals in our experiments
Dataset | Method | MOS |
|---|---|---|
UTE | Uniform sampling | 4.17±.30 |
|  | Clustering-based [4] | 5.10±.29 |
|  | Attention-based [8] | 4.49±.23 |
|  | Object-driven [2] | 5.36±.32 |
|  | Proposed (w/o optimization) | 5.18±.22 |
|  | Proposed | 5.6 9±.29 |
ADL | Uniform sampling | 4.41±.20 |
|  | Clustering-based [4] | 5.25±.18 |
|  | Attention-based [8] | 4.57±.16 |
|  | Proposed (w/o optimization) | 5.26±.14 |
|  | Proposed | 5.5 1±.16 |