Abstract
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: this https URL.
Abstract (translated)
近年来，随着自动3D虚拟角色生成技术基于文本的进步，已经取得了显著的进展。然而，现有的方法存在一些局限性，如过饱和和低质量输出。为了应对这些挑战，我们提出了X-Oscar，一种基于文本提示生成高质量动感的虚拟角色的渐进框架。它遵循了序列化几何->纹理->动画范式，通过逐步生成简化优化。为了解决过饱和问题，我们引入了自适应变分参数（AVP），将虚拟角色表示为在训练期间的自适应分布。此外，我们提出了Avatar-aware Score Distillation Sampling（ASDS），一种将虚拟意识到的噪声融入渲染图像中以提高优化过程中的生成质量的新技术。广泛的评估证实了X-Oscar在现有文本到3D和文本到虚拟角色方法中的优越性。我们的匿名项目页面：这是这个链接：https:// this URL。
URL
https://arxiv.org/abs/2405.00954
