Research map / Image generation / Editing and style transfer
A Temporal and Content Co-Awareness Latent Diffusion for Controllable Hand Image Generation
Shuang Hao, Pengfei Ren, Haifeng Sun, Ting Pan, Qi Qi, Lei Zhang, Cong Liu, Jianxin Liao, Jingyu Wang
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 the full TCCA module on top of PIAE (ablation) | TCCA (SD 1.5) · InterHand2.6M subset | FID: 11.089 -> 9.046 | Improved |
| add TCCA on the full FoundHand-10M (ablation) | TCCA (SD 1.5) · Full FoundHand-10M | FID: 9.647 -> 8.183 | Improved |
| use full TCCA instead of simple MLP fusion (ablation) | TCCA (SD 1.5) · InterHand2.6M subset | FID: 11.436 -> 9.046 | Improved |
| add content-awareness to coarse temporal cues (ablation) | TCCA · InterHand2.6M subset | FID: 10.263 -> 9.046 | Improved |
| add temporal-awareness to content-awareness only (ablation) | TCCA · InterHand2.6M subset | FID: 9.804 -> 9.046 | Improved |
| add content-awareness to temporal-awareness only (ablation) | TCCA · InterHand2.6M subset | FID: 9.652 -> 9.046 | Improved |
| add global CLS token to patch features in PIAE (ablation) | PIAE · InterHand2.6M subset | FID: 15.423 -> 11.089 | Improved |
| add patch-level features to CLS token in PIAE (ablation) | PIAE · InterHand2.6M subset | FID: 14.002 -> 11.089 | Improved |
| add SVD-based orthogonal decomposition to remove pose components (ablation) | PIAE · InterHand2.6M subset | FID: 12.946 -> 11.089 | Improved |
| use TCCA instead of FoundHand baseline | controllable hand generation · InterHand2.6M hand pose evaluation | PCK@0.05: 83.93 -> 87.34 | Improved |
| remove TCCA but keep PIAE versus FoundHand on full FoundHand-10M (ablation) | TCCA without TCCA module · Full FoundHand-10M | LPIPS: 0.4504 (FoundHand) -> 0.4632 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- FoundHand — FoundHand: Large-Scale Domain-Specific Learning for Controllable Hand Image Generation (CVPR 2025) · better on 8 of 8 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- FoundHand: Large-Scale Domain-Specific Learning for Controllable Hand Image Generation (CVPR 2025 · similarity 0.89)
- InterHandGen: Two-Hand Interaction Generation via Cascaded Reverse Diffusion (CVPR 2024 · similarity 0.88)
- HanDiffuser: Text-to-Image Generation With Realistic Hand Appearances (CVPR 2024 · similarity 0.87)
- Multi-focal Conditioned Latent Diffusion for Person Image Synthesis (CVPR 2025 · similarity 0.87)
- GeneOH Diffusion: Towards Generalizable Hand-Object Interaction Denoising via Denoising Diffusion (ICLR 2024 · similarity 0.86)
- SViMo: Synchronized Diffusion for Video and Motion Generation in Hand-object Interaction Scenarios (NeurIPS 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.
