Time-To-Collision forecasting on CCD
0.37MSECollideNet
Evaluation results
| Method | Links | |
|---|---|---|
| CollideNetBackbone=Multi-scale ViT2026.04 | 0.37 | |
| VidNeXtBackbone=ConvNeXt2026.04 | 0.53 | |
| VGG-16Backbone=VGG-162026.04 | 0.54 | |
| Video-FocalNetBackbone=Focal Mod. Net2026.04 | 0.58 | |
| Video Swin Trans.Backbone=Swin Transformer2026.04 | 0.62 | |
| TimeSformerBackbone=Vision Transformer2026.04 | 0.63 | |
| HyCTBackbone=CNN-Transformer2026.04 | 0.78 | |
| ViViTBackbone=Vision Transformer2026.04 | 0.82 | |
| X3DBackbone=3D ConvNet2026.04 | 0.85 | |
| ResNet50 3DBackbone=3D ConvNet2026.04 | 0.87 | |
| CNN-RNNBackbone=EfficientNet2026.04 | 0.93 | |
| Li3DBackbone=3D ConvNet2026.04 | 1.42 | |
| C3DBackbone=3D ConvNet2026.04 | 1.98 |