Research Article
Moving Vehicle Detection and Classification Using Gaussian Mixture Model and Ensemble Deep Learning Technique
Table 4
Performance analysis of the proposed ensemble deep learning technique on MIO-TCD by means of precision, recall, and accuracy.
| Feature extraction | Classifier | Precision (%) | Recall (%) | Accuracy (%) |
| SPT | MSVM | 67.84 | 70 | 82 | KNN | 70.40 | 70.27 | 65 | DNN | 78.22 | 89 | 80.10 | LSTM | 85 | 88.29 | 88 | Ensemble | 89 | 89.54 | 90.72 | WLD | MSVM | 70 | 80 | 90 | KNN | 78 | 80.90 | 92.04 | DNN | 83 | 90.72 | 93 | LSTM | 87.09 | 92 | 94.20 | Ensemble | 89 | 93.39 | 97 | Hybrid (SPT + WLD) | MSVM | 72.34 | 93.02 | 96 | KNN | 82.02 | 94 | 97.77 | DNN | 92.31 | 96.20 | 96 | LSTM | 98.70 | 98.82 | 98.62 | Ensemble | 99.12 | 99.69 | 99.13 |
|
|