Trajectory Forecasting on Zara 2 UCY scene (test)
0.11ADETrajectron++
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Trajectron++mode=probabilistic, output_type=Full, best_of_n_samples=202020.01 | 0.11 | 0.25 | |
| Trajectron++mode=probabilistic, output_type=Full, best_of_n_samples=20, integration_via_dynamics=true2020.01 | 0.12 | 0.25 | |
| Trajectron++mode=deterministic, output_type=Most Likely (ML)2020.01 | 0.23 | 0.59 | |
| Trajectron++mode=deterministic, output_type=Most Likely (ML), integration_via_dynamics=true2020.01 | 0.23 | 0.59 | |
| MATFmode=probabilistic, best_of_n_samples=202020.01 | 0.26 | 0.57 | |
| S-ATTNmode=deterministic2020.01 | 0.3 | 2.13 | |
| SoPhiemode=probabilistic, best_of_n_samples=202020.01 | 0.38 | 0.78 | |
| S-GANmode=probabilistic, best_of_n_samples=202020.01 | 0.42 | 0.84 | |
| Trajectronmode=probabilistic, best_of_n_samples=202020.01 | 0.43 | 0.85 | |
| LSTMmode=deterministic2020.01 | 0.52 | 1.11 | |
| S-LSTMmode=deterministic2020.01 | 0.56 | 1.17 | |
| Linearmode=deterministic2020.01 | 0.77 | 1.48 |