Research Article
A Study of Feature Combination for Vehicle Detection Based on Image Processing
Table 7
Performance of the proposed ensembles for the front close/middle region.
(a) |
| PCA+V-HOG +++ | Accuracy | Recall | Precision |
| RMM-SA | 99.26 | 99.84 | 99.68 | RMM-WA | 99.20 | 99.72 | 98.68 | DS-SA | 99.16 | 99.42 | 98.56 | DS-WA | 99.42 | 99.76 | 99.08 |
|
|
(b) |
| PCA+V-HOG + + | Accuracy | Recall | Precision |
| RMM-SA | 98.38 | 99.15 | 97.60 | RMM-WA | 98.74 | 99.51 | 97.96 | DS-SA | 98.38 | 99.15 | 97.60 | DS-WA | 99.12 | 99.64 | 98.60 |
|
|
(c) |
| PCA+V-HOG+ | Accuracy | Recall | Precision |
| RMM-SA | 99.02 | 99.60 | 98.44 | RMM-WA | 99.10 | 99.60 | 98.60 | DS-SA | 99.04 | 99.60 | 98.48 | DS-WA | 99.12 | 99.60 | 98.64 |
|
|
(d) |
| PCA+ | Accuracy | Recall | Precision |
| RMM-SA | 98.64 | 99.51 | 97.76 | RMM-WA | 98.88 | 99.39 | 98.36 | DS-SA | 98.58 | 99.51 | 97.64 | DS-WA | 98.94 | 99.39 | 98.48 |
|
|