Research Article
Multimodal Semantics Extraction from User-Generated Videos
Table 6
Experimental results on identifying the optimal cameras. P stands for Precision, R for Recall, and F for F-measure.
| Event | All cameras (index) | Ground truth sets of optimal cameras | Automatically identified optimal set of cameras | P | R | F |
| Football match 1 | | | | 0.8 | 1.0 | 0.89 | Football match 2 | | | | 1.0 | 1.0 | 1.0 | Football match 3 | | | | 0.75 | 1.0 | 0.86 | Concert 1 | | | | 1.0 | 1.0 | 1.0 | Concert 2 | | | | 1.0 | 1.0 | 1.0 |
| Average over all events | — | — | — | 0.91 | 1.0 | 0.95 |
|
|