Table 2 Speed and performances for trained model with the 2007 and 2012 VOC datasets
From: Object detection in real time based on improved single shot multi-box detector algorithm
Model | mAP | FPS | Real-time speed |
|---|---|---|---|
Faster YOLO | 52.7% | 155 | YES |
YOLO | 63.4% | 45 | YES |
YOLO VGG-16 | 66.4% | 21 | NO |
Fast R-CNN | 70.0% | 0.5 | NO |
Faster R-CNN VGG-16 | 73.2% | 7 | NO |
Faster R-CNN ZF | 62.1% | 18 | NO |