Research map / 3D vision / Human pose estimation
Collaborative Learning for Hand and Object Reconstruction With Attention-Guided Graph Convolution
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 collaborative hand-object reconstruction without explicit contact loss | Ours · FHB− | hand error (mm): 28.1 -> 25.3 | Helped |
| use collaborative hand-object reconstruction without explicit contact loss | Ours · DexYCB− | object error (mm): 549.4 -> 501.2 | Helped |
| add associative loss (ablation) | collaborative reconstruction at P=2 · FHB− | hand error: 26.3 -> 25.3 | Helped |
| add associative loss (ablation) | collaborative reconstruction at P=2 · FHB− | object error: 1618.4 -> 1445.0 | Helped |
| use attention-guided graph convolution instead of GCN (ablation) | collaborative reconstruction at P=2 with associative loss · FHB− | hand error: 27.0 -> 25.3 | Helped |
| use attention-guided graph convolution instead of Spiral convolution (ablation) | collaborative reconstruction at P=2 with associative loss · FHB− | hand error: 26.9 -> 25.3 | Helped |
| increase iterations from P=1 to P=2 (ablation) | Ours with associative loss · FHB− | hand error: 26.9 -> 25.3 | Helped |
| increase iterations from P=2 to P=3 (ablation) | Ours with associative loss · FHB− | hand error: 25.3 -> 25.4 | No clear effect |
| use associative loss (ablation) | Ours across network iterations · FHB− | training stability | Helped |
| share reconstructed mesh information within iterations instead of predicting embeddings and reconstructing only at the final stage (ablation) | collaborative learning framework · DexYCB S0 | hand error: 17.5 -> 16.1 | Helped |
| share reconstructed mesh information within iterations instead of predicting embeddings and reconstructing only at the final stage (ablation) | collaborative learning framework · FHB− | object error: 1726.8 -> 1445.0 | Helped |
| increase iterations for static GCN (ablation) | GCN collaborative reconstruction · FHB− | hand error: 27.1 -> 27.0 | Hurt |
Closest papers by meaning
- Interacting Attention Graph for Single Image Two-Hand Reconstruction (CVPR 2022 · similarity 0.91)
- ACR: Attention Collaboration-Based Regressor for Arbitrary Two-Hand Reconstruction (CVPR 2023 · similarity 0.90)
- Learning Explicit Contact for Implicit Reconstruction of Hand-Held Objects from Monocular Images (AAAI 2024 · similarity 0.90)
- Fine-Grained Multi-View Hand Reconstruction Using Inverse Rendering (AAAI 2024 · similarity 0.89)
- HOLD: Category-agnostic 3D Reconstruction of Interacting Hands and Objects from Video (CVPR 2024 · similarity 0.89)
- HandBooster: Boosting 3D Hand-Mesh Reconstruction by Conditional Synthesis and Sampling of Hand-Object Interactions (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.