Table 5 ME module ablation test results.
From: Improved 3D-ResNet sign language recognition algorithm with enhanced hand features
Method | FLOPs (G) | Parm (M) | Top1 (%) | Top2 (%) | Top3 (%) | Top4 (%) | Top5 (%) | AVG (%) |
|---|---|---|---|---|---|---|---|---|
Baseline | 8.308 | 33.246 | 81.06 | 88.98 | 92.22 | 94.00 | 94.78 | 90.21 |
Baseline + ME (Stage2) | 8.308 | 33.247 | 87.62 | 95.64 | 97.58 | 98.54 | 99.00 | 95.68 |
Baseline + ME (Stage3) | 8.308 | 33.249 | 88.42 | 95.60 | 97.32 | 98.22 | 98.70 | 95.65 |
Baseline + ME (Stage4) | 8.308 | 33.257 | 88.16 | 95.46 | 97.44 | 98.26 | 98.82 | 95.63 |
Baseline + ME (Stage5) | 8.308 | 33.289 | 88.64 | 95.78 | 97.72 | 98.52 | 98.94 | 95.92 |
Baseline + ME (Stage2–5) | 8.310 | 33.306 | 88.80 | 95.80 | 97.72 | 98.52 | 98.86 | 95.94 |
