Active 6D Multi-Object Pose Estimation in Cluttered Scenarios with Deep Reinforcement Learning

被引:10
|
作者
Sock, Juil [1 ]
Garcia-Hernando, Guillermo [1 ]
Kim, Tae-Kyun [1 ,2 ]
机构
[1] Imperial Coll London, Dept Elect & Elect Engn, London, England
[2] Korea Adv Inst Sci & Technol, Sch Comp, Daejeon, South Korea
关键词
D O I
10.1109/IROS45743.2020.9340842
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
In this work, we explore how a strategic selection of camera movements can facilitate the task of 6D multiobject pose estimation in cluttered scenarios while respecting real-world constraints such as time and distance travelled, important in robotics and augmented reality applications. In the proposed framework, multiple object hypotheses inferred by an object pose estimator are accumulated both in space and time with a fusion function. At each time step, this fusion function makes use of a verification score to quantify the quality of the hypotheses in the absence of ground-truth annotations and passes this information to an agent. The agent reasons about these hypotheses, directing its attention to the object which it is most uncertain about, moving the camera towards such an object. Unlike previous works that propose short-sighted policies, our agent is trained in simulated scenarios using reinforcement learning, attempting to learn the camera moves that produce the most accurate object poses hypotheses for a given temporal and spatial budget, without the need of viewpoints rendering during inference. Our experiments show that the proposed approach successfully estimates the 6D object pose of a stack of objects in both challenging cluttered synthetic and real scenarios, showing superior performance compared to other baselines.
引用
收藏
页码:10564 / 10571
页数:8
相关论文
共 50 条
  • [1] Coupled Iterative Refinement for 6D Multi-Object Pose Estimation
    Lipson, Lahav
    Teed, Zachary
    Goyal, Ankit
    Deng, Jia
    [J]. 2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2022), 2022, : 6718 - 6727
  • [2] Accurate 6D Object Pose Estimation and Refinement in Cluttered Scenes
    Jin, Yixiang
    Rossiter, John Anthony
    Veres, Sandor M.
    [J]. PROCEEDINGS OF THE 2ND INTERNATIONAL CONFERENCE ON ROBOTICS, COMPUTER VISION AND INTELLIGENT SYSTEMS (ROBOVIS), 2021, : 31 - 39
  • [3] Learning latent geometric consistency for 6D object pose estimation in heavily cluttered scenes
    Li, Qingnan
    Hu, Ruimin
    Xiao, Jing
    Wang, Zhongyuan
    Chen, Yu
    [J]. JOURNAL OF VISUAL COMMUNICATION AND IMAGE REPRESENTATION, 2020, 70
  • [4] 6D Object Pose Estimation in Cluttered Scenes from RGB Images
    Xiao-Long Yang
    Xiao-Hong Jia
    Yuan Liang
    Lu-Bin Fan
    [J]. Journal of Computer Science and Technology, 2022, 37 : 719 - 730
  • [5] 6D Object Pose Estimation in Cluttered Scenes from RGB Images
    Yang, Xiao-Long
    Jia, Xiao-Hong
    Liang, Yuan
    Fan, Lu-Bin
    [J]. JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY, 2022, 37 (03) : 719 - 730
  • [6] Deep instance segmentation and 6D object pose estimation in cluttered scenes for robotic autonomous grasping
    Wu, Yongxiang
    Fu, Yili
    Wang, Shuguo
    [J]. INDUSTRIAL ROBOT-THE INTERNATIONAL JOURNAL OF ROBOTICS RESEARCH AND APPLICATION, 2020, 47 (04): : 593 - 606
  • [7] YOLO-6D-Pose: Enhancing YOLO for Single-Stage Monocular Multi-Object 6D Pose Estimation
    Maji, Debapriya
    Nagori, Soyeb
    Mathew, Manu
    Poddar, Deepak
    [J]. 2024 INTERNATIONAL CONFERENCE IN 3D VISION, 3DV 2024, 2024, : 1616 - 1625
  • [8] PoseCNN: A Convolutional Neural Network for 6D Object Pose Estimation in Cluttered Scenes
    Xiang, Yu
    Schmidt, Tanner
    Narayanan, Venkatraman
    Fox, Dieter
    [J]. ROBOTICS: SCIENCE AND SYSTEMS XIV, 2018,
  • [9] PoseAgent: Budget-Constrained 6D Object Pose Estimation via Reinforcement Learning
    Krull, Alexander
    Brachmann, Eric
    Nowozin, Sebastian
    Michel, Frank
    Shotton, Jamie
    Rother, Carsten
    [J]. 30TH IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2017), 2017, : 2566 - 2574
  • [10] YOLOPose: Transformer-Based Multi-object 6D Pose Estimation Using Keypoint Regression
    Amini, Arash
    Periyasamy, Arul Selvam
    Behnke, Sven
    [J]. INTELLIGENT AUTONOMOUS SYSTEMS 17, IAS-17, 2023, 577 : 392 - 406