Researchers have developed AvatarDynamizer, a novel generative method designed to transform static 3D avatars into dynamic, realistic, and multi-view-consistent 4D avatars. This approach embeds pose-dependent dynamics into texture maps, allowing compatibility with pre-trained video diffusion models and rendering them using 3D Gaussians. To address limitations in existing datasets, a large-scale multi-view dataset with diverse motions and surface dynamics was collected, demonstrating AvatarDynamizer's superior visual fidelity, particularly with limited dynamic training data. AI
IMPACT This research could enhance realism in virtual environments and digital human representation.
RANK_REASON The item is an academic paper detailing a new method for avatar generation. [lever_c_demoted from research: ic=1 ai=1.0]
- 3D avatars
- 3D Gaussians
- 4D avatar
- alphaXiv
- arXiv
- AvatarDynamizer
- CatalyzeX Code Finder for Papers
- Connected Papers
- CORE Recommender
- DagsHub
- Dynamic Textures Modeling via Joint Video Dictionary Learning.
- Gotit.pub
- Hugging Face
- Influence Flower
- Litmaps
- ScienceCast
- scite Smart Citations
- Video Diffusion Models
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
