Research map / 3D vision / Human pose estimation
Head2Body: Body Pose Generation from Multi-sensory Head-mounted Inputs
Minh Tran, Hongda Mao, Qingshuang Chen, Yelin Kim
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 vector quantization of poses (ablation) | VQ-Poser · AMASS | MPJPE (mm): 104.3 -> 90.5 | Improved |
| add final pose predictor Ppose (ablation) | VQ-Poser · AMASS | MPJPE (mm): 99.2 -> 90.5 | Improved |
| use VQ-Poser instead of transformer encoder with linear regression (ablation) | Head2Body · AMASS | MPJPE (mm): 114.9 -> 90.5 | Improved |
| add head-IMU input (ablation) | Head2Body · AMASS | MPJPE (mm): 100.9 -> 98 | Improved |
| pre-train IMU encoder on Ego4D (ablation) | Head2Body · AMASS | MPJPE (mm): 98 -> 90.5 | Improved |
| use our pre-trained IMU encoder instead of IMU2CLIP (ablation) | Head2Body · AMASS | MPJPE (mm): 93.2 -> 90.5 | Improved |
| add visual body-part input (ablation) | Head2Body · EgoExo4D test | MPJPE (cm): 14.3 -> 13.0 | Improved |
| add body part segmentation before image encoder (ablation) | Head2Body · EgoExo4D test | MPJPE (cm): 13.8 -> 13.0 | Improved |
| add visual input on procedural activities (ablation) | Head2Body · EgoExo4D test (procedural) | Pr-MPJPE (cm): 9.9 -> 8.1 | Improved |
| add visual input on physical activities (ablation) | Head2Body · EgoExo4D test (physical) | Ph-MPJPE (cm): 15.3 -> 14.1 | Improved |
| use estimated head pose from HeadNet instead of ground truth (ablation) | Head2Body · ARES | MPJPE (mm): 86.9 -> 107.9 | Hurt |
| use Head2Body instead of MocapEvery with estimated head pose | Head2Body · ARES | MPJPE (mm): 128.7 -> 107.9 | Improved |
Closest papers by meaning
- FRAME: Floor-aligned Representation for Avatar Motion from Egocentric Video (CVPR 2025 · similarity 0.91)
- Estimating Ego-Body Pose from Doubly Sparse Egocentric Video Data (NeurIPS 2024 · similarity 0.90)
- EgoSim: An Egocentric Multi-view Simulator and Real Dataset for Body-worn Cameras during Motion and Activity (NeurIPS 2024 · similarity 0.90)
- EMHI: A Multimodal Egocentric Human Motion Dataset with HMD and Body-Worn IMUs (AAAI 2025 · similarity 0.89)
- Real-Time Simulated Avatar from Head-Mounted Sensors (CVPR 2024 · similarity 0.88)
- Estimating Body and Hand Motion in an Ego-sensed World (CVPR 2025 · similarity 0.88)
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.
