基于自注意网络的非裁剪RGB图像双手姿态估计

Zhoutao Sun, Yong Hu, Xukun Shen
{"title":"基于自注意网络的非裁剪RGB图像双手姿态估计","authors":"Zhoutao Sun, Yong Hu, Xukun Shen","doi":"10.1109/ismar52148.2021.00040","DOIUrl":null,"url":null,"abstract":"Estimating the pose of two hands is a crucial problem for many human-computer interaction applications. Since most of the existing works utilize cropped images to predict the hand pose, they require a hand detection stage before pose estimation or input cropped images directly. In this paper, we propose the first real-time one-stage method for pose estimation from a single RGB image without hand tracking. Combining the self-attention mechanism with convolutional layers, the network we proposed is able to predict the 2.5D hand joints coordinate while locating the two hands regions. And to reduce the extra memory and computational consumption caused by self-attention, we proposed a linear attention structure with a spatial reduction attention block called SRAN block. We demonstrate the effectiveness of each component in our network through the ablation study. And experiments on public datasets showed the competitive result with the state-of-the-art method.","PeriodicalId":395413,"journal":{"name":"2021 IEEE International Symposium on Mixed and Augmented Reality (ISMAR)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":"{\"title\":\"Two-hand Pose Estimation from the non-cropped RGB Image with Self-Attention Based Network\",\"authors\":\"Zhoutao Sun, Yong Hu, Xukun Shen\",\"doi\":\"10.1109/ismar52148.2021.00040\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Estimating the pose of two hands is a crucial problem for many human-computer interaction applications. Since most of the existing works utilize cropped images to predict the hand pose, they require a hand detection stage before pose estimation or input cropped images directly. In this paper, we propose the first real-time one-stage method for pose estimation from a single RGB image without hand tracking. Combining the self-attention mechanism with convolutional layers, the network we proposed is able to predict the 2.5D hand joints coordinate while locating the two hands regions. And to reduce the extra memory and computational consumption caused by self-attention, we proposed a linear attention structure with a spatial reduction attention block called SRAN block. We demonstrate the effectiveness of each component in our network through the ablation study. And experiments on public datasets showed the competitive result with the state-of-the-art method.\",\"PeriodicalId\":395413,\"journal\":{\"name\":\"2021 IEEE International Symposium on Mixed and Augmented Reality (ISMAR)\",\"volume\":\"15 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2021-10-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"1\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2021 IEEE International Symposium on Mixed and Augmented Reality (ISMAR)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ismar52148.2021.00040\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 IEEE International Symposium on Mixed and Augmented Reality (ISMAR)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ismar52148.2021.00040","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1

摘要

估计双手的姿态是许多人机交互应用中的一个关键问题。由于现有的工作大多是利用裁剪后的图像来预测手部姿态,因此在姿态估计之前需要一个手部检测阶段,或者直接输入裁剪后的图像。在本文中,我们提出了第一个实时的单阶段方法,用于从单个RGB图像中进行姿态估计,而无需手动跟踪。将自注意机制与卷积层相结合,我们提出的网络能够在定位双手区域的同时预测2.5D手部关节坐标。为了减少自注意带来的额外内存和计算消耗,我们提出了一种具有空间缩减注意块的线性注意结构,称为SRAN块。我们通过烧蚀研究证明了网络中每个组件的有效性。在公共数据集上的实验表明,该方法具有较强的竞争力。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Two-hand Pose Estimation from the non-cropped RGB Image with Self-Attention Based Network
Estimating the pose of two hands is a crucial problem for many human-computer interaction applications. Since most of the existing works utilize cropped images to predict the hand pose, they require a hand detection stage before pose estimation or input cropped images directly. In this paper, we propose the first real-time one-stage method for pose estimation from a single RGB image without hand tracking. Combining the self-attention mechanism with convolutional layers, the network we proposed is able to predict the 2.5D hand joints coordinate while locating the two hands regions. And to reduce the extra memory and computational consumption caused by self-attention, we proposed a linear attention structure with a spatial reduction attention block called SRAN block. We demonstrate the effectiveness of each component in our network through the ablation study. And experiments on public datasets showed the competitive result with the state-of-the-art method.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
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