Table 4 Ablation experiment results.
From: Hybrid-YOLOv5 for object detection of non-ferrous metals in end-of-life vehicles
Baseline | C2 F | SE | MobileNetv3 | P(%) | R(%) | F1 (%)score | mAP_0.5(%) | FLOPs (G) | Infer time (ms) | FPS(Frames/s) |
|---|---|---|---|---|---|---|---|---|---|---|
√ | 71.9 | 80.7 | 76.1 | 74.9 | 16 | 21.5 | 43 | |||
√ | √ | 72.5 | 85.1 | 78.3 | 80.3 | 73.8 | 21.2 | 44 | ||
√ | √ | 72.0 | 83.4 | 77.3 | 78.1 | 48.4 | 15.6 | 58 | ||
√ | √ | 71.0 | 77.9 | 74.3 | 74.6 | 10.7 | 15.5 | 59 | ||
√ | √ | √ | 72.4 | 80.0 | 76.0 | 79.3 | 74.9 | 21.5 | 43 | |
√ | √ | √ | 69.4 | 80.9 | 74.7 | 74.9 | 12.8 | 15.4 | 59 | |
√ | √ | √ | 66.0 | 75.2 | 70.3 | 74.6 | 10.7 | 17.9 | 50 | |
√ | √ | √ | √ | 72.3 | 79.9 | 75.9 | 84.2 | 12.8 | 15.2 | 60 |
