Spatiotemporal Predictive Learning on Moving MNIST 10 time steps 3-digit (test)
0.814SSIMPredRNN++
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PredRNN++description=Causal LSTM + GHU (Final)2018.04 | 0.814 | 81.7 | |
| Causal LSTM2018.04 | 0.795 | 89.2 | |
| PredRNN + GHU2018.04 | 0.79 | 88.9 | |
| Causal LSTMvariant=spatial-to-temporal2018.04 | 0.784 | 91.8 | |
| PredRNN2018.04 | 0.782 | 93.4 | |
| VPNreproduced=true2018.04 | 0.734 | 112.3 | |
| TrajGRU2018.04 | 0.682 | 134 | |
| DFN2018.04 | 0.679 | 140.5 | |
| ConvLSTM2018.04 | 0.673 | 142.1 | |
| CDNA2018.04 | 0.669 | 138.2 | |
| FC-LSTM2018.04 | 0.651 | 162.4 |