Table 6 Comparison of the recognition effects of various model files for similar fire and smoke
Method | False alarm rate (%) | Accuracy (%) |
|---|---|---|
Types of recognition | Fire | |
Faster R-CNN (Ren et al. 2015) | 33.1% | 67.1% |
SSD (Liu et al. 2016) | 27.3% | 68.9% |
YOLOv3 (Redmon et al. 2018) | 31.2% | 68.9% |
YOLOv4 | 13.1% | 74.3% |
MobileNetV3-large-YOLOv4 | 11.1% | 85.8% |
Types of recognition | Smoke | |
Faster R-CNN | 35.6% | 63.7% |
SSD | 33.6% | 62.9% |
YOLOv3 | 35.3% | 64.7% |
YOLOv4 | 16.7% | 74.5% |
MobileNetV3-large-YOLOv4 | 15.8% | 76.3% |