用于 6DoF 物体姿态估计的端到端概率几何引导回归技术

Thomas Pöllabauer, Jiayin Li, Volker Knauthe, Sarah Berkei, Arjan Kuijper
{"title":"用于 6DoF 物体姿态估计的端到端概率几何引导回归技术","authors":"Thomas Pöllabauer, Jiayin Li, Volker Knauthe, Sarah Berkei, Arjan Kuijper","doi":"arxiv-2409.11819","DOIUrl":null,"url":null,"abstract":"6D object pose estimation is the problem of identifying the position and\norientation of an object relative to a chosen coordinate system, which is a\ncore technology for modern XR applications. State-of-the-art 6D object pose\nestimators directly predict an object pose given an object observation. Due to\nthe ill-posed nature of the pose estimation problem, where multiple different\nposes can correspond to a single observation, generating additional plausible\nestimates per observation can be valuable. To address this, we reformulate the\nstate-of-the-art algorithm GDRNPP and introduce EPRO-GDR (End-to-End\nProbabilistic Geometry-Guided Regression). Instead of predicting a single pose\nper detection, we estimate a probability density distribution of the pose.\nUsing the evaluation procedure defined by the BOP (Benchmark for 6D Object Pose\nEstimation) Challenge, we test our approach on four of its core datasets and\ndemonstrate superior quantitative results for EPRO-GDR on LM-O, YCB-V, and\nITODD. Our probabilistic solution shows that predicting a pose distribution\ninstead of a single pose can improve state-of-the-art single-view pose\nestimation while providing the additional benefit of being able to sample\nmultiple meaningful pose candidates.","PeriodicalId":501130,"journal":{"name":"arXiv - CS - Computer Vision and Pattern Recognition","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"End-to-End Probabilistic Geometry-Guided Regression for 6DoF Object Pose Estimation\",\"authors\":\"Thomas Pöllabauer, Jiayin Li, Volker Knauthe, Sarah Berkei, Arjan Kuijper\",\"doi\":\"arxiv-2409.11819\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"6D object pose estimation is the problem of identifying the position and\\norientation of an object relative to a chosen coordinate system, which is a\\ncore technology for modern XR applications. State-of-the-art 6D object pose\\nestimators directly predict an object pose given an object observation. Due to\\nthe ill-posed nature of the pose estimation problem, where multiple different\\nposes can correspond to a single observation, generating additional plausible\\nestimates per observation can be valuable. To address this, we reformulate the\\nstate-of-the-art algorithm GDRNPP and introduce EPRO-GDR (End-to-End\\nProbabilistic Geometry-Guided Regression). Instead of predicting a single pose\\nper detection, we estimate a probability density distribution of the pose.\\nUsing the evaluation procedure defined by the BOP (Benchmark for 6D Object Pose\\nEstimation) Challenge, we test our approach on four of its core datasets and\\ndemonstrate superior quantitative results for EPRO-GDR on LM-O, YCB-V, and\\nITODD. Our probabilistic solution shows that predicting a pose distribution\\ninstead of a single pose can improve state-of-the-art single-view pose\\nestimation while providing the additional benefit of being able to sample\\nmultiple meaningful pose candidates.\",\"PeriodicalId\":501130,\"journal\":{\"name\":\"arXiv - CS - Computer Vision and Pattern Recognition\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-18\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - CS - Computer Vision and Pattern Recognition\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.11819\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - CS - Computer Vision and Pattern Recognition","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.11819","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

6D 物体姿态估计是识别物体相对于选定坐标系的位置和方向的问题,是现代 XR 应用的核心技术。最先进的 6D 物体姿态估计器可在观测到物体的情况下直接预测物体姿态。由于姿态估计问题的多假设性质,一个观测值可能对应多个不同的姿态,因此为每个观测值生成额外的可信估计值非常有价值。为了解决这个问题,我们重新制定了最先进的算法 GDRNPP,并引入了 EPRO-GDR(端到端概率几何引导回归)。使用 BOP(6D 物体姿态估计基准)挑战赛定义的评估程序,我们在其四个核心数据集上测试了我们的方法,并在 LM-O、YCB-V 和 ITODD 上展示了 EPRO-GDR 优越的定量结果。我们的概率解决方案表明,预测姿态分布而非单一姿态可以改进最先进的单视角姿态估算,同时还能提供对多个有意义的候选姿态进行采样的额外好处。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
End-to-End Probabilistic Geometry-Guided Regression for 6DoF Object Pose Estimation
6D object pose estimation is the problem of identifying the position and orientation of an object relative to a chosen coordinate system, which is a core technology for modern XR applications. State-of-the-art 6D object pose estimators directly predict an object pose given an object observation. Due to the ill-posed nature of the pose estimation problem, where multiple different poses can correspond to a single observation, generating additional plausible estimates per observation can be valuable. To address this, we reformulate the state-of-the-art algorithm GDRNPP and introduce EPRO-GDR (End-to-End Probabilistic Geometry-Guided Regression). Instead of predicting a single pose per detection, we estimate a probability density distribution of the pose. Using the evaluation procedure defined by the BOP (Benchmark for 6D Object Pose Estimation) Challenge, we test our approach on four of its core datasets and demonstrate superior quantitative results for EPRO-GDR on LM-O, YCB-V, and ITODD. Our probabilistic solution shows that predicting a pose distribution instead of a single pose can improve state-of-the-art single-view pose estimation while providing the additional benefit of being able to sample multiple meaningful pose candidates.
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:481959085
Book学术官方微信