Research map / 3D vision / Human pose estimation
Human Shape From Silhouettes Using Generative HKS Descriptors and Cross-Modal Neural Networks
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 the proposed SFUS-HKS-1 method instead of HS-Net-1-S | Human shape estimation from a scaled frontal silhouette · Synthetic body measurements, one scaled frontal silhouette | mean error (mm): 10.8 mm -> 4.01 mm | Helped |
| Use the proposed UF-US-HKS-2 method instead of HS-2-Net-MM | Human shape estimation from two views · Synthetic body measurements, two views | mean error (mm): 10.8 mm -> 3.77 mm | Helped |
| Add a side-view silhouette during training (ablation) | CMNN single-view prediction · Synthetic body measurements, one scaled frontal silhouette | mean error | Helped |
| Add an unscaled frontal view during training (ablation) | CMNN single-view prediction · Synthetic body measurements, one scaled frontal silhouette | mean error | Helped |
| Train with scaled frontal, unscaled frontal, and scaled side views (ablation) | CMNN single-view prediction · Synthetic body measurements, one scaled frontal silhouette | mean error | Helped |
| Jointly train HKS-Net with CMNN rather than training HKS-Net separately (ablation) | HKS-Net shape embedding · HKS-Net mesh reconstruction | mean error (mm): 4.74 mm -> 3.77 mm | Helped |
| Train and test with different poses (ablation) | SF-1-P · Synthetic body measurements, one scaled frontal silhouette | mean error | No clear effect |
| Add high levels of noise to silhouettes | Proposed method · Silhouette noise test | mean error increase (mm): +4.1 mm | Hurt |
Closest papers by meaning
- Detailed Human Shape Estimation From a Single Image by Hierarchical Mesh Deformation (CVPR 2019 · similarity 0.90)
- SiCloPe: Silhouette-Based Clothed People (CVPR 2019 · similarity 0.89)
- View Extrapolation of Human Body From a Single Image (CVPR 2018 · similarity 0.89)
- SCULPT: Shape-Conditioned Unpaired Learning of Pose-dependent Clothed and Textured Human Meshes (CVPR 2024 · similarity 0.86)
- Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimation (NeurIPS 2024 · similarity 0.86)
- MUC: Mixture of Uncalibrated Cameras for Robust 3D Human Body Reconstruction (AAAI 2025 · similarity 0.86)
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.