Table 8 The influence of each module on the performance evaluation indicators of the model in ablation experiments.
From: Object detection model design for tiny road surface damage
Backbone | Neck | MSD-head | mAP50(%) | F1 | Params(M) | GFLOPs | |||
|---|---|---|---|---|---|---|---|---|---|
DSCD | MFE | SPPF | FFM | FSM | |||||
✓ | 65.3 | 65.2 | 21.4 | 28.4 | |||||
✓ | ✓ | 66.4 | 66.5 | 20.0 | 26.8 | ||||
✓ | ✓ | 66.5 | 66.9 | 20.3 | 26.0 | ||||
✓ | ✓ | 65.2 | 65.5 | 17.6 | 23.9 | ||||
✓ | ✓ | ✓ | 66.9 | 67.0 | 18.8 | 24.4 | |||
✓ | ✓ | 65.8 | 65.6 | 19.8 | 26.8 | ||||
✓ | ✓ | ✓ | 66.8 | 66.5 | 20.0 | 27.8 | |||
✓ | ✓ | 69.0 | 68.7 | 20.6 | 36.4 | ||||
✓ | ✓ | ✓ | ✓ | ✓ | 68.3 | 67.5 | 17.4 | 23.7 | |
✓ | ✓ | ✓ | ✓ | 69.7 | 69.0 | 18.0 | 32.4 | ||
✓ | ✓ | ✓ | ✓ | ✓ | ✓ | 70.8 | 70.2 | 16.5 | 30.9 |
