Research map / 3D vision / Human pose estimation
Deep Multitask Architecture for Integrated 2D and 3D Human Sensing
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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-part maps B and joint maps J to the D-based reconstruction pipeline (ablation) | DMHS_R on Human80K · Human80K | Avg. MPJPE (mm): 77.56 -> 63.35 | Helped |
| Add body-part maps B to the J,D reconstruction input (ablation) | DMHS_R on Human80K · Human80K | Avg. MPJPE (mm): 72.00 -> 63.35 | Helped |
| Add image features D to the J-only reconstruction input (ablation) | DMHS_R on Human80K · Human80K | Avg. MPJPE (mm): 128.05 -> 72.00 | Helped |
| Add body-part maps B to J-only reconstruction (ablation) | DMHS_R on Human80K · Human80K | Avg. MPJPE (mm): 128.05 -> 118.68 | Hurt |
| Train body-part segmentation jointly on Human80K and LSP instead of Human80K alone (ablation) | DMHS_B · Human80K test | Avg. Acc. per class (fg + bg) (%): 68.56 -> 66.13 | Hurt |
| Fine-tune the Human80K segmentation model on LSP (ablation) | DMHS_B · Human80K test | Avg. Acc. per class (fg + bg) (%): 68.56 -> 45.61 | Hurt |
| Train body-part segmentation jointly on Human80K and LSP rather than Human80K alone (ablation) | DMHS_B · LSP | Avg. Acc. per class (fg + bg) (%): 38.77 -> 48.01 | Helped |
| Fine-tune the Human80K segmentation model on LSP (ablation) | DMHS_B · LSP | Avg. Acc. per class (fg + bg) (%): 38.77 -> 46.88 | Helped |
Closest papers by meaning
- Monocular 3D Pose and Shape Estimation of Multiple People in Natural Scenes - The Importance of Multiple Scene Constraints (CVPR 2018 · similarity 0.89)
- 2D/3D Pose Estimation and Action Recognition Using Multitask Deep Learning (CVPR 2018 · similarity 0.89)
- Deep Network for the Integrated 3D Sensing of Multiple People in Natural Images (NeurIPS 2018 · similarity 0.89)
- DeepCut: Joint Subset Partition and Labeling for Multi Person Pose Estimation (CVPR 2016 · similarity 0.87)
- ManiPose: Manifold-Constrained Multi-Hypothesis 3D Human Pose Estimation (NeurIPS 2024 · similarity 0.85)
- FreeMan: Towards Benchmarking 3D Human Pose Estimation under Real-World Conditions (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.