Research map / 3D vision / Human pose estimation
Coarse-To-Fine Volumetric Prediction for Single-Image 3D Human Pose
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 |
|---|---|---|---|
| use volumetric likelihood prediction instead of coordinate regression (ablation) | single-hourglass pose estimator · Human3.6M | mean per joint error (mm): 112.41 -> 85.82 | Helped |
| increase volume depth resolution from d = 32 to d = 64 (ablation) | Volume Regression · Human3.6M | mean per joint error (mm): 92.23 -> 85.82 | Helped |
| add coarse-to-fine depth supervision rather than naively stacking two hourglasses (ablation) | two-hourglass pose estimator · Human3.6M | mean per joint error (mm): 80.14 -> 69.77 | Helped |
| add coarse-to-fine depth supervision rather than naively stacking three hourglasses (ablation) | three-hourglass pose estimator · Human3.6M | mean per joint error (mm): 78.17 -> 68.49 | Helped |
| add coarse-to-fine depth supervision rather than naively stacking four hourglasses (ablation) | four-hourglass pose estimator · Human3.6M | mean per joint error (mm): 75.06 -> 64.76 | Helped |
| use an intermediate depth resolution of 8 instead of 1 (ablation) | two-stage Coarse-to-Fine · Human3.6M | mean per joint error (mm): 69.77 -> 77.52 | Hurt |
| add an intermediate depth resolution of 2 before 64 (ablation) | three-stage Coarse-to-Fine · Human3.6M | mean per joint error (mm): 72.02 -> 68.49 | Helped |
| use depth supervision resolutions 1, 2, 4, then 64 instead of 1, 2, then 64 (ablation) | Coarse-to-Fine · Human3.6M | mean per joint error (mm): 68.49 -> 64.76 | Helped |
| retain image features in the 3D prediction stages rather than using 2D heatmaps alone (ablation) | pose estimator · Human3.6M | mean per joint error (mm): 78.10 -> 69.77 | Helped |
Closest papers by meaning
- Learning to Estimate 3D Human Pose and Shape From a Single Color Image (CVPR 2018 · similarity 0.93)
- 3D Human Pose Estimation = 2D Pose Estimation + Matching (CVPR 2017 · similarity 0.92)
- Lifting From the Deep: Convolutional 3D Pose Estimation From a Single Image (CVPR 2017 · similarity 0.92)
- 3D Human Pose Estimation From a Single Image via Distance Matrix Regression (CVPR 2017 · similarity 0.91)
- Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimation (NeurIPS 2024 · similarity 0.89)
- PandaPose: 3D Human Pose Lifting from a Single Image via Propagating 2D Pose Prior to 3D Anchor Space (NeurIPS 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.