High-Fidelity Clothed Avatar Reconstruction from a Single Image

Tingting LiaoXiaomei ZhangYuliang XiuHongwei YiXudong LiuGuo-Jun QiYong ZhangXuan WangXiangyu ZhuZhen Lei

Tingting LiaoXiaomei ZhangYuliang Xiu ...+6 Zhen Lei

Apr 2023

摘要

This paper presents a framework for efficient 3D clothed avatarreconstruction. By combining the advantages of the high accuracy ofoptimization-based methods and the efficiency of learning-based methods, wepropose a coarse-to-fine way to realize a high-fidelity clothed avatarreconstruction (CAR) from a single image. At the first stage, we use animplicit model to learn the general shape in the canonical space of a person ina learning-based way, and at the second stage, we refine the surface detail byestimating the non-rigid deformation in the posed space in an optimization way.A hyper-network is utilized to generate a good initialization so that theconvergence o f the optimization process is greatly accelerated. Extensiveexperiments on various datasets show that the proposed CAR successfullyproduces high-fidelity avatars for arbitrarily clothed humans in real scenes.

展开全部

图表提取

论文十问

参考文献

被引用

提问与回答

暂无人提供速读十问回答

被引用

笔记

问答

领域