X-Oscar: A Progressive Framework for High-quality Text-guided 3D Animatable Avatar Generation
This work addresses limitations in 3D avatar generation for applications like gaming or virtual reality, representing an incremental improvement with novel techniques.
The paper tackles the problem of oversaturation and low-quality output in text-guided 3D animatable avatar generation by proposing X-Oscar, a progressive framework that achieves superior performance over existing methods in evaluations.
Recent advancements in automatic 3D avatar generation guided by text have made significant progress. However, existing methods have limitations such as oversaturation and low-quality output. To address these challenges, we propose X-Oscar, a progressive framework for generating high-quality animatable avatars from text prompts. It follows a sequential Geometry->Texture->Animation paradigm, simplifying optimization through step-by-step generation. To tackle oversaturation, we introduce Adaptive Variational Parameter (AVP), representing avatars as an adaptive distribution during training. Additionally, we present Avatar-aware Score Distillation Sampling (ASDS), a novel technique that incorporates avatar-aware noise into rendered images for improved generation quality during optimization. Extensive evaluations confirm the superiority of X-Oscar over existing text-to-3D and text-to-avatar approaches. Our anonymous project page: https://xmu-xiaoma666.github.io/Projects/X-Oscar/.