研究文章
卷积神经网络分类和决策模型可见缺陷高速列车图像的识别
|
| 选项 |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
|
| 包括(180360)、(360180)的形状吗? |
✓ |
✓ |
✓ |
✓ |
✓ |
✓ |
✓ |
✓ |
| 包括(148442)、(442148)的形状吗? |
|
|
|
✓ |
✓ |
✓ |
|
|
| 1-shape推理吗? |
|
✓ |
|
|
|
|
|
|
| 这种理想状态推理吗? |
✓ |
|
|
✓ |
|
|
✓ |
|
| 3-shape推理吗? |
|
|
✓ |
|
|
✓ |
|
✓ |
| 5-shape推理吗? |
|
|
|
|
✓ |
|
|
|
| 翻转吗? |
|
|
|
|
|
|
✓ |
✓ |
| 精度(%) |
93.67 |
92.90 |
94.05 |
93.67 |
94.00 |
94.08 |
94.38 |
95.08 |
| 精度(%) |
94.44 |
93.16 |
95.41 |
95.17 |
95.18 |
95.27 |
94.71 |
95.73 |
| 回忆(%) |
92.76 |
92.60 |
92.55 |
92.00 |
92.71 |
92.75 |
94.60 |
94.21 |
| F1 (%) |
93.59 |
92.88 |
93.96 |
93.56 |
93.92 |
94.00 |
94.35 |
94.96 |
|
|
