Research Article
Object Extraction of Tennis Video Based on Deep Learning
Table 1
Algorithm performance comparison.
| | Category | VOC2007mAP (%) | FPS (titan X) | Number of generated boxes | Enter picture resolution |
| | R-CNN | 65.7 | 5 | ~6000 | ~1000600 | | Faster R-CNN | 74.1 | 7 | ~6000 | ~1000600 | | HyperNet | 80.6 | 45 | 128 | 512512 |
|
|
