Table 12 The impact of different attention mechanisms on model performance
From: GrainNet: efficient detection and counting of wheat grains based on an improved YOLOv7 modeling
| Â | Â | mAP@0.5(%) | mAP@0.5:0.95(%) | Params(M) | FLOP(G) |
|---|---|---|---|---|---|
I | Baseline + Lightweight + EMA | 89.22 | 56.18 | 26.76 | 71.7 |
II | Baseline + Lightweight + ECA | 86.66 | 55.69 | 26.78 | 71.6 |
III | Baseline + Lightweight + CBAM | 87.60 | 55.93 | 26.81 | 71.6 |
IV | Baseline + Lightweight + SE | 86.81 | 55.61 | 26.77 | 71.6 |