Efficient 6D object pose estimation based on attentive multi-scale contextual information | |
Gao, Fang1; Sun, Qingyi1; Li, Shaodong1; Li, Wenbo2; Li, Yong1; Yu, Jun3; Shuang, Feng1 | |
刊名 | IET COMPUTER VISION |
2022-04-02 | |
ISSN号 | 1751-9632 |
DOI | 10.1049/cvi2.12101 |
通讯作者 | Li, Shaodong(lishaodongyx@126.com) ; Li, Yong(yongli@gxu.edu.cn) |
英文摘要 | 6D pose estimation has been pervasively applied to various robotic applications, such as service robots, collaborative robots, and unmanned warehouses. However, accurate 6D pose estimation is still a challenge problem due to the complexity of application scenarios caused by illumination changes, occlusion and even truncation between objects, and additional refinement is required for accurate 6D object pose estimation in prior work. Aiming at the efficiency and accuracy of 6D object pose estimation in these complex scenes, this paper presents a novel end-to-end network, which effectively utilises the contextual information within a neighbourhood region of each pixel to estimate the 6D object pose from RGB-D images. Specifically, our network first applies the attention mechanism to extract effective pixel-wise dense multimodal features, which are then expanded to multi-scale dense features by integrating pixel-wise features at different scales for pose estimation. The proposed method is evaluated extensively on the LineMOD and YCB-Video datasets, and the experimental results show that the proposed method is superior to several state-of-the-art baselines in terms of average point distance and average closest point distance. |
资助项目 | National Natural Science Foundation of China[41871302] ; National Natural Science Foundation of China[61773359] ; National Natural Science Foundation of China[61720106009] ; Guangxi Key Laboratory of Manufacturing System & Advanced Manufacturing Technology[20-065-40S005] ; Guangxi Science and Technology base and Talent Project[2020AC19253] ; USTC Research Funds of the Double First-Class Initiative[YD2350002001] ; Anhui Provincial Natural Science Foundation[2108085J19] ; Anhui Province Key Research and Development Program[202104a05020007] ; CAAI-Huawei MindSpore Open Fund[CAAIXSJLJJ-2021-016B] |
WOS研究方向 | Computer Science ; Engineering |
语种 | 英语 |
出版者 | WILEY |
WOS记录号 | WOS:000777139300001 |
资助机构 | National Natural Science Foundation of China ; Guangxi Key Laboratory of Manufacturing System & Advanced Manufacturing Technology ; Guangxi Science and Technology base and Talent Project ; USTC Research Funds of the Double First-Class Initiative ; Anhui Provincial Natural Science Foundation ; Anhui Province Key Research and Development Program ; CAAI-Huawei MindSpore Open Fund |
内容类型 | 期刊论文 |
源URL | [http://ir.hfcas.ac.cn:8080/handle/334002/128679] |
专题 | 中国科学院合肥物质科学研究院 |
通讯作者 | Li, Shaodong; Li, Yong |
作者单位 | 1.Guangxi Univ, Guangxi Key Lab Intelligent Control & Maintenance, 100 Daxue East Rd, Nanning 530004, Peoples R China 2.Chinese Acad Sci, Inst Intelligent Machines, Hefei, Anhui, Peoples R China 3.Univ Sci & Technol China, Dept Automat, Hefei, Anhui, Peoples R China |
推荐引用方式 GB/T 7714 | Gao, Fang,Sun, Qingyi,Li, Shaodong,et al. Efficient 6D object pose estimation based on attentive multi-scale contextual information[J]. IET COMPUTER VISION,2022. |
APA | Gao, Fang.,Sun, Qingyi.,Li, Shaodong.,Li, Wenbo.,Li, Yong.,...&Shuang, Feng.(2022).Efficient 6D object pose estimation based on attentive multi-scale contextual information.IET COMPUTER VISION. |
MLA | Gao, Fang,et al."Efficient 6D object pose estimation based on attentive multi-scale contextual information".IET COMPUTER VISION (2022). |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论