Table 2 Comparison of prediction results of different prediction models.
From: Research on imaging method of driver's attention area based on deep neural network
FPS | Unit | 1 | .… | 39 | 40 | .… | 80 | 81 | … | 120 | MSE |
|---|---|---|---|---|---|---|---|---|---|---|---|
True value | Angle | 0.000 | .… | − 5.000 | − 5.000 | .… | − 4.000 | − 4.000 | … | − 3.000 | 0.000 |
Speed | 33.000 | 35.000 | 36.000 | 0.000 | 0.000 | 0.000 | 0.000 | ||||
This modle | Angle | − 0.213 | − 3.872 | − 4.187 | − 4.843 | − 2.668 | − 4.986 | 0.362 | |||
Speed | 32.097 | 35.792 | 35.449 | − 0.390 | 0.120 | 1.248 | 0.224 | ||||
Inattentive | Angle | 3.150 | − 4.218 | − 8.068 | − 0.468 | − 4.368 | 1.024 | 0.568 | |||
Speed | 34.229 | 33.603 | 32.333 | − 1.470 | 0.656 | 3.293 | 0.623 | ||||
Single-layer GRU | Angle | 1.946 | − 8.173 | − 2.157 | − 3.092 | − 3.231 | − 5.953 | 2.543 | |||
Speed | 35.111 | 36.662 | 37.532 | − 3.345 | − 2.654 | − 0.783 | 2.896 | ||||
LSTM | Angle | 0.794 | − 5.626 | − 0.078 | 2.017 | − 2.977 | − 4.455 | 5.426 | |||
Speed | 31.183 | 39.346 | 31.105 | 5.698 | − 3.068 | − 5.794 | 5.867 |
