License Plate Recognition on AOLP (full)
97.59AccuracyOurs
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| OursAblation=None (Full Objective)2019.10 | 97.59 | 95.89 | 96.87 | 96.74 | |
| Ours (without pixel-wise MSE loss)Ablation=without pixel-wise MSE loss2019.10 | 97.24 | 94.67 | 94.91 | 95.6 | |
| Ours (without classification loss)Ablation=without classification loss2019.10 | 96.39 | 94.98 | 96.48 | 95.88 | |
| Ours (without reconstruction loss)Ablation=without reconstruction loss2019.10 | 96.21 | 88.89 | 94.32 | 92.91 | |
| Smith2019.10 | 96 | 83 | 83 | 87.31 | |
| Ours (without adversarial loss)Ablation=without adversarial loss2019.10 | 95.18 | 87.67 | 93.93 | 92 | |
| Hsu et al.2019.10 | 95 | 93 | 94 | 94.17 | |
| Baseline (YOLO v3)Architecture=YOLO v32019.10 | 94.66 | 89.04 | 89.04 | 90.9 | |
| Anagnostopoulos et al.2019.10 | 92 | 88 | 91 | 86.34 | |
| Jiao et al.2019.10 | 90 | 86 | 90 | 88.51 |