Research Article
Encoding Longer-Term Contextual Information with Predictive Coding and Ego-Motion
Table 3
RMS between LSTM and MT-AFA-PredNet.
| ā | LSTM | MT-AFA-PredNet |
| Construction site | 8.332 | 7.219 |
| Crazy turn | 10.315 | 8.287 |
| Dancing light | 9.834 | 7.314 |
| Intern on bike | 5.411 | 5.131 |
| Safe turn | 9.314 | 8.107 |
| Squirrel | 7.908 | 7.781 |
|
|