A Method to Generate Posed Person Image with few Context Images

H. Nakada, H. Asoh
{"title":"A Method to Generate Posed Person Image with few Context Images","authors":"H. Nakada, H. Asoh","doi":"10.1109/imcom53663.2022.9721635","DOIUrl":null,"url":null,"abstract":"We report a method to generate an image of an arbitrary person in an arbitrary pose, where the person’s characteristics are provided as a few sample images. The rapid development in image generation techniques using deep neural networks makes it possible to generate photo-realistic images. However, controlling the content of the generated images, such as a person’s clothing or pose, is still an open problem. The method we propose creates a pose invariant person representation using the attention mechanism and generates posed images by applying pose query on the representation. We evaluated the method using 3D rendered synthetic data and confirmed that the created person representation is pose-invariant, and we can render good quality images with the representation.","PeriodicalId":367038,"journal":{"name":"2022 16th International Conference on Ubiquitous Information Management and Communication (IMCOM)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 16th International Conference on Ubiquitous Information Management and Communication (IMCOM)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/imcom53663.2022.9721635","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

Abstract

We report a method to generate an image of an arbitrary person in an arbitrary pose, where the person’s characteristics are provided as a few sample images. The rapid development in image generation techniques using deep neural networks makes it possible to generate photo-realistic images. However, controlling the content of the generated images, such as a person’s clothing or pose, is still an open problem. The method we propose creates a pose invariant person representation using the attention mechanism and generates posed images by applying pose query on the representation. We evaluated the method using 3D rendered synthetic data and confirmed that the created person representation is pose-invariant, and we can render good quality images with the representation.
一种基于少量背景图像生成人物姿态图像的方法
我们报告了一种生成任意姿势的任意人物图像的方法,其中人物的特征作为几个样本图像提供。基于深度神经网络的图像生成技术的快速发展使得生成逼真的图像成为可能。然而,控制生成图像的内容,比如一个人的服装或姿势,仍然是一个悬而未决的问题。我们提出的方法利用注意机制创建姿态不变的人表示,并在该表示上应用姿态查询生成姿态图像。我们使用3D渲染合成数据对该方法进行了评估,并证实了所创建的人物表示是位姿不变的,并且我们可以使用该表示来渲染高质量的图像。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信