Research map / 3D vision / Human pose estimation
Monocular Expressive Body Regression through Body-Driven Attention
Vasileios Choutas, Georgios Pavlakos, Timo Bolkart, Dimitrios Tzionas, Michael J. Black
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 body-driven attention with independently trained hand/face sub-networks (ablation) | Body only network · EHF | PA-V2V all (mm): 57.3 -> 56.4 | Improved |
| add body-driven attention with independently trained hand/face sub-networks (ablation) | Body only network · EHF | PA-V2V body (mm): 55.9 -> 52.6 | Improved |
| add body-driven attention with independently trained hand/face sub-networks (ablation) | Body only network · EHF | PA-V2V face (mm): 5.8 -> 6.0 | Hurt |
| add end-to-end fine-tuning of body, hand and face networks (ablation) | Body & Hand & Face attention · EHF | PA-V2V all (mm): 56.4 -> 54.5 | Improved |
| add end-to-end fine-tuning of body, hand and face networks (ablation) | Body & Hand & Face attention · EHF | PA-V2V L/R hand (mm): 14.1 / 13.9 -> 13.1 / 12.5 | Improved |
| add end-to-end fine-tuning of body, hand and face networks (ablation) | Body & Hand & Face attention · EHF | PA-V2V body (mm): 52.6 -> 52.6 | No clear effect |
| use ExPose instead of body-only low-resolution regression (ablation) | Body only network · EHF | PA-V2V all (mm): 57.3 -> 54.5 | Improved |
| use ExPose regression instead of SPIN (ablation) | SPIN · 3DPW | PA-MPJPE (mm): 59.2 -> 60.7 | No clear effect |
| use ExPose instead of HMR | HMR · 3DPW | PA-MPJPE (mm): 81.3 -> 60.7 | Improved |
| use ExPose regression instead of SMPLify-X optimization | SMPLify-X · EHF | PA-V2V all (mm): 65.3 -> 54.5 | Improved |
| use ExPose regression instead of SMPLify-X optimization | SMPLify-X · EHF | runtime (s): 40-60 -> 0.16 | Improved |
| use ExPose hand sub-network instead of MANO CNN | MANO CNN · FreiHAND | PA-MPJPE (mm): 11.0 -> 12.2 | Hurt |
Closest papers by meaning
- Expressive Body Capture: 3D Hands, Face, and Body From a Single Image (CVPR 2019 · similarity 0.95)
- SUPR: A Sparse Unified Part-Based Human Representation (ECCV 2022 · similarity 0.90)
- Monocular Real-Time Full Body Capture With Inter-Part Correlations (CVPR 2021 · similarity 0.90)
- Expressive Gaussian Human Avatars from Monocular RGB Video (NeurIPS 2024 · similarity 0.90)
- MUC: Mixture of Uncalibrated Cameras for Robust 3D Human Body Reconstruction (AAAI 2025 · similarity 0.88)
- AiOS: All-in-One-Stage Expressive Human Pose and Shape Estimation (CVPR 2024 · 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.
