Research Article
Predicting Pedestrian Counts for Crossing Scenario Based on Fused Infrared-Visual Videos
Table 6
The results of 8-fold cross validation.
| The number of test | Visible light video only | Infrared video only | Fused information |
| 1 | 66.67% | 73.33% | 80.00% |
| 2 | 80.00% | 86.67% | 93.33% |
| 3 | 73.33% | 66.67% | 80.00% |
| 4 | 81.25% | 75.00% | 87.50% |
| 5 | 66.67% | 80.00% | 86.67% |
| 6 | 73.33% | 66.67% | 80.00% |
| 7 | 80.00% | 80.00% | 86.67% |
| 8 | 68.75% | 87.50% | 87.50% |
| Average accuracy | 73.75% | 76.98% | 85.21% |
|
|