Text2Face: Text-Based Face Generation With Geometry and Appearance Control

doi:10.1109/TVCG.2023.3349050

	Text2Face: Text-Based Face Generation With Geometry and Appearance Control
	Zhang, Zhaoyang 1; Chen, Junliang 4; Fu, Hongbo 5; Zhao, Jianjun 4; Chen, Shu-Yu 2,3; Gao, Lin 2,3
	2024-09-01
发表期刊	IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS
ISSN	1077-2626
卷号	30 期号:9 页码:6481-6492
摘要	Recent years have witnessed the emergence of various techniques proposed for text-based human face generation and manipulation. Such methods, targeting bridging the semantic gap between text and visual contents, provide users with a deft hand to turn ideas into visuals via text interface and enable more diversified multimedia applications. However, due to the flexibility of linguistic expressiveness, the mapping from sentences to desired facial images is clearly many-to-many, causing ambiguities during text-to-face generation. To alleviate these ambiguities, we introduce a local-to-global framework with two graph neural networks (one for geometry and the other for appearance) embedded to model the inter-dependency among facial parts. This is based upon our key observation that the geometry and appearance attributes among different facial components are not mutually independent, i.e., the combinations of part-level facial features are not arbitrary and thus do not conform to a uniform distribution. By learning from the dataset distribution and enabling recommendations given partial descriptions of human faces, these networks are highly suitable for our text-to-face task. Our method is capable of generating high-quality attribute-conditioned facial images from text. Extensive experiments have confirmed the superiority and usability of our method over the prior art.
关键词	Faces Geometry Feature extraction Semantics Pipelines Nose Image synthesis Image generation face editing sketching interface text-based user interaction
DOI	10.1109/TVCG.2023.3349050
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[62061136007] ; National Natural Science Foundation of China[62102403] ; Beijing Municipal Natural Science Foundation for Distinguished Young Scholars[JQ21013] ; Beijing Municipal Science and Technology Commission[Z231100005923031]
WOS研究方向	Computer Science
WOS类目	Computer Science, Software Engineering
WOS记录号	WOS:001283711000023
出版者	IEEE COMPUTER SOC
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/39658
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Gao, Lin
作者单位	1.Yale Univ, Dept Comp Sci, New Haven, CT 06520 USA 2.Chinese Acad Sci, Inst Comp Technol, Beijing Key Lab Mobile Comp & Pervas Device, Beijing 100190, Peoples R China 3.Univ Chinese Acad Sci, Beijing 100190, Peoples R China 4.Beijing Film Acad, Dept Film & TV Technol, Beijing 100088, Peoples R China 5.City Univ Hong Kong, Sch Creat Media, Kowloon, Hong Kong, Peoples R China
推荐引用方式 GB/T 7714	Zhang, Zhaoyang,Chen, Junliang,Fu, Hongbo,et al. Text2Face: Text-Based Face Generation With Geometry and Appearance Control[J]. IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS,2024,30(9):6481-6492.
APA	Zhang, Zhaoyang,Chen, Junliang,Fu, Hongbo,Zhao, Jianjun,Chen, Shu-Yu,&Gao, Lin.(2024).Text2Face: Text-Based Face Generation With Geometry and Appearance Control.IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS,30(9),6481-6492.
MLA	Zhang, Zhaoyang,et al."Text2Face: Text-Based Face Generation With Geometry and Appearance Control".IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS 30.9(2024):6481-6492.