Pedestrian Intention Prediction on JAAD Beh
73AccuracyESIA
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ESIAYear=2026, Blocks=Att+ViT+CRF, Inputs=B,S,I, Input Frames=162026.04 | 73 | 0.7 | 79 | 73 | 87 | |
| PedCMTYear=2024, Blocks=Att, Inputs=B,S, Input Frames=162026.04 | 71 | 0.64 | 80 | — | — | |
| PedGraph+Year=2022, Blocks=GCN+CNN, Inputs=S,P,I, Input Frames=322026.04 | 70 | 0.7 | 76 | 77 | 75 | |
| PITYear=2023, Blocks=Att, Inputs=P,S,I, Input Frames=162026.04 | 70 | 0.65 | 81 | 71 | 93 | |
| PPCI_attYear=2024, Blocks=LSTM+Att, Inputs=B,S,P, Input Frames=162026.04 | 67 | 0.6 | 77 | — | — | |
| PedSAYear=2025, Blocks=Att+ViT, Inputs=B,S,P,I, Input Frames=162026.04 | 67 | 0.6 | 77 | 68 | 90 | |
| LSOP-NetYear=2025, Blocks=GRU+Att+CNN, Inputs=B,S,P,I, Input Frames=162026.04 | 65 | 0.54 | 78 | 65 | 98 | |
| Ours 2modality=non-visual, model_size=small2026.01 | 64 | 0.58 | 77 | 65 | 93 | |
| TrouSPI-NetYear=2021, Blocks=GRU+Att, Inputs=B,S,P, Input Frames=162026.04 | 64 | 0.56 | 76 | 66 | 91 | |
| Ours 1model_size=small2026.01 | 62 | 0.61 | 77 | 63 | 99 | |
| Ours 3model_size=large2026.01 | 62 | 0.55 | 73 | 66 | 83 | |
| FFSTPYear=2022, Blocks=CNN+GRU+Att, Inputs=B,S,P,I, Input Frames=162026.04 | 62 | 0.54 | 74 | 65 | 85 | |
| ConvLSTM2026.01 | 59 | 0.53 | 70 | 64 | 77 | |
| C3D2026.01 | 58 | 0.48 | 73 | 61 | 91 | |
| Single-GRUYear=2020, Blocks=GRU, Inputs=B,S,I, Input Frames=162026.04 | 58 | 0.54 | 67 | 67 | 68 | |
| PCPAYear=2021, Blocks=CNN+RNN+Att, Inputs=B,S,P,I, Input Frames=162026.04 | 58 | 0.5 | 71 | — | — | |
| PCPA2026.01 | 54 | 0.49 | 65 | 62 | 68 | |
| CAPformer2026.01 | 51 | 0.54 | 52 | 68 | 42 | |
| Single-LSTMYear=2020, Blocks=LSTM, Inputs=B,S,I, Input Frames=162026.04 | 51 | 0.48 | 61 | 63 | 59 |