Research Article
The Automatic Detection of Pedestrians under the High-Density Conditions by Deep Learning Techniques
Table 2
The statistics when detecting the 4 typical samples by our model.
| ā | T1 | T2 | T3 | T4 |
| (a) The results of Figure 3(a) | Precision | 0.96 | 1.00 | 1.00 | 0.99 | Recall | 0.16 | 0.89 | 0.99 | 1.00 | F1 score | 0.27 | 0.94 | 1.00 | 0.99 |
| (b) The results of Figure 3(b) | Precision | 0.86 | 0.94 | 0.97 | 0.98 | Recall | 0.14 | 0.88 | 0.97 | 1.00 | F1 score | 0.25 | 0.91 | 0.97 | 0.99 |
| (c) The results of Figure 3(c) | Precision | 0.98 | 1.00 | 1.00 | 0.99 | Recall | 0.10 | 0.79 | 0.87 | 0.98 | F1 score | 0.17 | 0.88 | 0.93 | 0.99 |
| (d) The results of Figure 3(d) | Precision | 0.95 | 0.99 | 0.99 | 0.99 | Recall | 0.08 | 0.64 | 0.81 | 0.95 | F1 score | 0.15 | 0.78 | 0.89 | 0.97 |
|
|