X-Oscar: A Progressive Framework for High-quality Text-guided 3D Animatable Avatar Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ma, Yiwei, Lin, Zhekai, Ji, Jiayi, Fan, Yijun, Sun, Xiaoshuai, Ji, Rongrong
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913339352809472
author Ma, Yiwei
Lin, Zhekai
Ji, Jiayi
Fan, Yijun
Sun, Xiaoshuai
Ji, Rongrong
author_facet Ma, Yiwei
Lin, Zhekai
Ji, Jiayi
Fan, Yijun
Sun, Xiaoshuai
Ji, Rongrong
contents 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/.
format Preprint
id arxiv_https___arxiv_org_abs_2405_00954
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle X-Oscar: A Progressive Framework for High-quality Text-guided 3D Animatable Avatar Generation
Ma, Yiwei
Lin, Zhekai
Ji, Jiayi
Fan, Yijun
Sun, Xiaoshuai
Ji, Rongrong
Computer Vision and Pattern Recognition
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/.
title X-Oscar: A Progressive Framework for High-quality Text-guided 3D Animatable Avatar Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2405.00954