Research map / 3D vision / Human pose estimation
Cambrian-P: Pose-Grounded Video Understanding
Jihan YANG, Zifan Zhao, Xichen Pan, Shusheng Yang, Junyi Zhang, Hu Xu, Shang-Wen Li, Saining Xie
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 |
|---|---|---|---|
| add camera pose supervision (pose tokens + head) | Cambrian-S fine-tuned on VSI-590K · VSI-Bench | accuracy: 69.2 -> 73.7 | Improved |
| add pose supervision (ablation) | Cambrian-P · VSI-Bench (32 frames) | accuracy: 67.3 -> 72.0 | Improved |
| add pose supervision at 128 frames (ablation) | Cambrian-P · VSI-Bench (128 frames) | accuracy: 71.2 -> 73.7 | Improved |
| add interleaved training (ablation) | Cambrian-P · VSI-Bench | accuracy: 69.4 -> 72.0 | Improved |
| add random jitter (ablation) | Cambrian-P · VSI-Bench | accuracy: 71.2 -> 72.0 | Improved |
| add pose supervision (ablation) | Cambrian-P · VSTI-Bench Cam. Mov. Dir. | accuracy: 67.7 -> 87.7 | Improved |
| add pose supervision (ablation) | Cambrian-P · Tomato (OOD) | accuracy: 20.4 -> 26.7 | Improved |
| add pose supervision (ablation) | Cambrian-P · MVBench (OOD) | accuracy: 51.9 -> 53.5 | Improved |
| add pseudo-pose supervision from in-the-wild video | Cambrian-P with VSI-590K + CamS-590K · EgoSchema | accuracy: 71.7 -> 73.6 | Improved |
| add pose conditioning at inference (ablation) | Cambrian-P · VSI-Bench | accuracy: 72.0 -> 72.0 | No clear effect |
| replace pose loss with depth loss (ablation) | Cambrian-P · VSI-Bench | accuracy: 72.0 -> 69.4 | Hurt |
| add depth loss to pose loss (ablation) | Cambrian-P · VSI-Bench | accuracy: 72.0 -> 71.7 | Hurt |
Closest papers by meaning
- Towards Understanding Camera Motions in Any Video (NeurIPS 2025 · similarity 0.86)
- Video-of-Thought: Step-by-Step Video Reasoning from Perception to Cognition (ICML 2024 · similarity 0.86)
- Motion-Grounded Video Reasoning: Understanding and Perceiving Motion at Pixel Level (CVPR 2025 · similarity 0.85)
- MVVBench: Benchmarking 4D Reasoning in Vision-Language Models (NeurIPS 2026 · similarity 0.85)
- SpinBench: Perspective and Rotation as a Lens on Spatial Reasoning in VLMs (ICLR 2026 · similarity 0.85)
- Coarse Correspondences Boost Spatial-Temporal Reasoning in Multimodal Language Model (CVPR 2025 · 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.
