Research map / Video / Video segmentation and tracking
Spatial-temporal Causal Inference for Partial Image-to-video Adaptation
Jin Chen, Xinxiao Wu, Yao Hu, Jiebo Luo
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 |
|---|---|---|---|
| Use the spatial-temporal causal framework with causality-guided mappings and class-wise alignment | partial image-to-video adaptation · S→U | classification accuracy (%): 84.63 -> 95.15 | Improved |
| Use the spatial-temporal causal framework with causality-guided mappings and class-wise alignment | partial image-to-video adaptation · E→H | classification accuracy (%): 32.80 -> 52.66 | Improved |
| Add effects inferred by causal inference (ablation) | image-to-video adaptation · S→U | classification accuracy (%): 90.74 -> 95.15 | Improved |
| Add effects inferred by causal inference (ablation) | image-to-video adaptation · E→H | classification accuracy (%): 43.79 -> 52.66 | Improved |
| Add the spatial domain shift effect (ablation) | image-to-video adaptation · S→U | classification accuracy (%): 91.92 -> 95.15 | Improved |
| Add the spatial domain shift effect (ablation) | image-to-video adaptation · E→H | classification accuracy (%): 48.40 -> 52.66 | Improved |
| Add the temporal domain shift effect (ablation) | image-to-video adaptation · S→U | classification accuracy (%): 93.85 -> 95.15 | Improved |
| Add the temporal domain shift effect (ablation) | image-to-video adaptation · E→H | classification accuracy (%): 45.33 -> 52.66 | Improved |
| Add the video-to-image mapping (ablation) | bidirectional heterogeneous image-video mappings · S→U | classification accuracy (%): 43.13 -> 95.15 | Improved |
| Add the video-to-image mapping (ablation) | bidirectional heterogeneous image-video mappings · E→H | classification accuracy (%): 38.05 -> 52.66 | Improved |
| Add cycle consistency loss (ablation) | bidirectional heterogeneous image-video mappings · S→U | classification accuracy (%): 34.80 -> 95.15 | Improved |
| Add cycle consistency loss (ablation) | bidirectional heterogeneous image-video mappings · E→H | classification accuracy (%): 32.96 -> 52.66 | Improved |
Closest papers by meaning
- Towards Uncertainty-aware Unsupervised Domain Adaptation for Videos and Time-Series with Causal Optimal Transport (CVPR 2026 · similarity 0.89)
- Where Do We (Not) Need Temporal Context in Low-Resource Video Task Adaptation? (NeurIPS 2026 · similarity 0.86)
- Visual Representation Learning through Causal Intervention for Controllable Image Editing (CVPR 2025 · similarity 0.86)
- Causal-Inspired Multitask Learning for Video-Based Human Pose Estimation (AAAI 2025 · similarity 0.86)
- Image-to-video Adaptation with Outlier Modeling and Robust Self-learning (AAAI 2025 · similarity 0.85)
- Modality-Collaborative Test-Time Adaptation for Action Recognition (CVPR 2024 · 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.
