Research map / Video / Video segmentation and tracking
Video Task Decathlon: Unifying Image and Video Tasks in Autonomous Driving
Thomas E. Huang, Yifan Liu, Luc Van Gool, Fisher Yu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| naive joint multi-task training without CPF | ResNet-50 multi-task baseline · VTD validation | VTDA: 218.2 -> 216.9 | Hurt |
| add CPF training protocol to multi-task baseline | ResNet-50 multi-task baseline · VTD validation | VTDA: 216.9 -> 221.5 | Improved |
| add CPF training protocol to multi-task baseline | Swin-T multi-task baseline · VTD validation | VTDA: 217.0 -> 223.7 | Improved |
| add feature interaction blocks (VTDNet) to multi-task baseline | ResNet-50 · VTD validation | VTDA: 221.5 -> 225.3 | Improved |
| add feature interaction blocks (VTDNet) to multi-task baseline | Swin-T · VTD validation | VTDA: 223.7 -> 226.9 | Improved |
| add Intra-IB (ablation) | VTDNet ResNet-50 · VTD validation | VTDA: 221.5 -> 223.9 | Improved |
| add Cross-IB on top of Intra-IB (ablation) | VTDNet ResNet-50 · VTD validation | VTDA: 223.9 -> 225.3 | Improved |
| add curriculum training (ablation) | VTDNet ResNet-50 · VTD validation | VTDA: 216.7 -> 217.8 | Improved |
| add pseudo-labels for pose and semantic segmentation (ablation) | VTDNet ResNet-50 · VTD validation | semantic segmentation IoU: 42.0 -> 63.2 | Improved |
| add fine-tuning (ablation) | VTDNet ResNet-50 · VTD validation | VTDA: 222.4 -> 225.3 | Improved |
| double loss weights of one task group (ablation) | VTDNet ResNet-50 · VTD validation | VTDA: 225.3 -> 224.7 | No clear effect |
| scale backbone from ResNet-50 to ConvNeXt-B | VTDNet · VTD validation | VTDA: 225.3 -> 230.7 | Improved |
Closest papers by meaning
- M3Net: Multimodal Multi-task Learning for 3D Detection, Segmentation, and Occupancy Prediction in Autonomous Driving (AAAI 2025 · similarity 0.88)
- DualAD: Disentangling the Dynamic and Static World for End-to-End Driving (CVPR 2024 · similarity 0.87)
- FIVE-VLA: Fast and EffectIVE Closed-Loop Autonomous Driving with Recurrent Action Memory (NeurIPS 2026 · similarity 0.85)
- Prune2Drive: A Plug-and-Play Framework for Accelerating Vision-Language Models in Autonomous Driving (CVPR 2026 · similarity 0.85)
- Impromptu VLA: Open Weights and Open Data for Driving Vision-Language-Action Models (NeurIPS 2025 · similarity 0.85)
- Grounding Driving VLA via Inverse Kinematics (NeurIPS 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
