Research map / Video / Video generation
GenLCA: 3D Diffusion for Full-Body Avatars from In-the-Wild Videos
Yiqian Wu, Rawal Khirodkar, Egor Zakharov, Timur Bagautdinov, Lei Xiao, Zhaoen Su, Shunsuke Saito, Xiaogang Jin, Junxuan Li
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 visibility-aware training (masked loss and placeholders over valid tokens) (ablation) | GenLCA · generated avatars | qualitative blur/transparency | Improved |
| use learnable placeholder tokens instead of fixed zero tokens (ablation) | GenLCA · generated avatars | qualitative color | Improved |
| add in-the-wild video data to indoor capture data (3,000 identities only) (ablation) | GenLCA · text-aligned generation | qualitative generalization | Improved |
| GenLCA vs SIGMAN on image-conditioned generation | GenLCA · THuman 2.0 | FID: 121.40 -> 96.03 | Improved |
| GenLCA vs SIGMAN on image-conditioned generation | GenLCA · HuGe100K | FID: 160.48 -> 76.50 | Improved |
| GenLCA vs TeRA on FID against 2D diffusion images | GenLCA · 2D-diffusion FID | FID: 151.80 -> 160.91 | Hurt |
| GenLCA vs DreamWaltz-G on user-study visual quality | GenLCA · user study | visual quality: 3.41 -> 4.65 | Improved |
| GenLCA vs DreamWaltz-G on semantic alignment (BLIP-VQA) | GenLCA · text-to-avatar | BLIP-VQA: 0.58 -> 0.64 | Improved |
| GenLCA vs SIGMAN on inference time | GenLCA · text-to-avatar | inference time: 3s -> 12s | Hurt |
Closest papers by meaning
- Morphable Diffusion: 3D-Consistent Diffusion for Single-image Avatar Creation (CVPR 2024 · similarity 0.92)
- Human-3Diffusion: Realistic Avatar Creation via Explicit 3D Consistent Diffusion Models (NeurIPS 2024 · similarity 0.90)
- DiffusionGAN3D: Boosting Text-guided 3D Generation and Domain Adaptation by Combining 3D GANs and Diffusion Priors (CVPR 2024 · similarity 0.89)
- VLOGGER: Multimodal Diffusion for Embodied Avatar Synthesis (CVPR 2025 · similarity 0.89)
- Zero-1-to-A: Zero-Shot One Image to Animatable Head Avatars Using Video Diffusion (CVPR 2025 · similarity 0.88)
- Make-Your-Anchor: A Diffusion-based 2D Avatar Generation Framework (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.
