Real-time 6D pose estimation from a single RGB image |
| |
| Affiliation: | 1. Department of Electrical and Computer Engineering and ASRI, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul 08826, Korea;2. Division of Electrical Engineering, Hanyang University, 55 Hanyangdaehak-ro, Sangnok-Gu, Ansan, Gyeonggi-do 15588, Korea;1. State Key Lab for Novel Software Technology, Nanjing University, Nanjing, China;2. Science and Technology on Information Systems Engineering Laboratory, Nanjing, China;1. University of Science and Technology of China, Hefei 230026, PR China;2. Peking University, Beijing 100000, PR China |
| |
| Abstract: | We propose an end-to-end deep learning architecture for simultaneously detecting objects and recovering 6D poses in an RGB image. Concretely, we extend the 2D detection pipeline with a pose estimation module to indirectly regress the image coordinates of the object's 3D vertices based on 2D detection results. Then the object's 6D pose can be estimated using a Perspective-n-Point algorithm without any post-refinements. Moreover, we elaborately design a backbone structure to maintain spatial resolution of low level features for pose estimation task. Compared with state-of-the-art RGB based pose estimation methods, our approach achieves competitive or superior performance on two benchmark datasets at an inference speed of 25 fps on a GTX 1080Ti GPU, which is capable of real-time processing. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
