DreamWaltz

library paper
IDEA Lab
Text-driven 3D animatable avatar generation using ControlNet and SMPL. Proposes 3D-consistent occlusion-aware Score Distillation Sampling for high-quality avatar creation without artifacts. Supports complex avatar-avatar, avatar-object, and avatar-scene interactions.
Paper (arXiv)GitHubProject Page

Outputs 2

DreamWaltz

library

Framework for text-driven 3D animatable avatar generation with skeleton-guided score distillation.

GitHubProject Page

GitHub Repository

DreamWaltz: Make a Scene with Complex 3D Animatable Avatars

paper

Learning framework for text-driven 3D animatable avatar creation using 3D-consistent occlusion-aware SDS and skeleton conditioning.

Paper (arXiv)

arXiv: 2305.12529

Venue: NeurIPS 2023

vision3dgenerationopen-source