Generative Category-Level Shape and Pose Estimation with Semantic Primitives

被引:0
|
作者
Li, Guanglin [1 ,2 ,4 ]
Li, Yifeng [2 ]
Ye, Zhichao [1 ]
Zhang, Qihang [3 ]
Kong, Tao [2 ]
Cui, Zhaopeng [1 ]
Zhang, Guofeng [1 ]
机构
[1] Zhejiang Univ, State Key Lab CAD&CG, Hangzhou, Zhejiang, Peoples R China
[2] Chinese Univ Hong Kong, ByteDance AI Lab, Hong Kong, Peoples R China
[3] Chinese Univ Hong Kong, Multimedia Lab, Hong Kong, Peoples R China
[4] ByteDance, Hong Kong, Peoples R China
来源
关键词
Category-level Pose Estimation; Shape Estimation;
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Empowering autonomous agents with 3D understanding for daily objects is a grand challenge in robotics applications. When exploring in an unknown environment, existing methods for object pose estimation are still not satisfactory due to the diversity of object shapes. In this paper, we propose a novel framework for category-level object shape and pose estimation from a single RGB-D image. To handle the intra-category variation, we adopt a semantic primitive representation that encodes diverse shapes into a unified latent space, which is the key to establish reliable correspondences between observed point clouds and estimated shapes. Then, by using a SIM(3)-invariant shape descriptor, we gracefully decouple the shape and pose of an object, thus supporting latent shape optimization of target objects in arbitrary poses. Extensive experiments show that the proposed method achieves SOTA pose estimation performance and better generalization in the real-world dataset. Code and video are available at https://zju3dv.github.io/gCasp.
引用
收藏
页码:1390 / 1400
页数:11
相关论文
共 50 条
  • [21] Toward Real-World Category-Level Articulation Pose Estimation
    Liu, Liu
    Xue, Han
    Xu, Wenqiang
    Fu, Haoyuan
    Lu, Cewu
    IEEE TRANSACTIONS ON IMAGE PROCESSING, 2022, 31 : 1072 - 1083
  • [22] Category-Level 6-D Object Pose Estimation With Shape Deformation for Robotic Grasp Detection
    Yu, Sheng
    Zhai, Di-Hua
    Guan, Yuyin
    Xia, Yuanqing
    IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2025, 36 (01) : 1857 - 1871
  • [23] Optimal and Robust Category-Level Perception: Object Pose and Shape Estimation From 2-D and 3-D Semantic Keypoints
    Shi, Jingnan
    Yang, Heng
    Carlone, Luca
    IEEE TRANSACTIONS ON ROBOTICS, 2023, 39 (05) : 4131 - 4151
  • [24] GarmentTracking: Category-Level Garment Pose Tracking
    Xue, Han
    Xu, Wenqiang
    Zhang, Jieyi
    Tang, Tutian
    Li, Yutong
    Du, Wenxin
    Ye, Ruolin
    Lu, Cewu
    2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2023, : 21233 - 21242
  • [25] HS-Pose: Hybrid Scope Feature Extraction for Category-level Object Pose Estimation
    Zheng, Linfang
    Wang, Chen
    Sun, Yinghan
    Dasgupta, Esha
    Chen, Hua
    Leonardis, Ales
    Zhang, Wei
    Chang, Hyung Jin
    2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2023, : 17163 - 17173
  • [26] Category-Level Object Pose Estimation in Heavily Cluttered Scenes by Generalized Two-Stage Shape Reconstructor
    Tatemichi, Hiroki
    Kawanishi, Yasutomo
    Deguchi, Daisuke
    Ide, Ichiro
    Murase, Hiroshi
    IEEE ACCESS, 2024, 12 : 33440 - 33448
  • [27] An efficient network for category-level 6D object pose estimation
    Sun, Shantong
    Liu, Rongke
    Sun, Shuqiao
    Yang, Xinxin
    Lu, Guangshan
    SIGNAL IMAGE AND VIDEO PROCESSING, 2021, 15 (07) : 1643 - 1651
  • [28] Sca-pose: category-level 6D pose estimation with adaptive shape prior based on CNN and graph convolution
    Zuo, Guoyu
    Yu, Shan
    Yu, Shuangyue
    Liu, Hong
    Zhao, Min
    INTELLIGENT SERVICE ROBOTICS, 2025, 18 (02) : 351 - 361
  • [29] CatFormer: Category-Level 6D Object Pose Estimation with Transformer
    Yu, Sheng
    Zhai, Di-Hua
    Xia, Yuanqing
    THIRTY-EIGHTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, VOL 38 NO 7, 2024, : 6808 - 6816
  • [30] RANSAC Optimization for Category-level 6D Object Pose Estimation
    Chen, Ying
    Kang, Guixia
    Wang, Yiping
    2020 5TH INTERNATIONAL CONFERENCE ON MECHANICAL, CONTROL AND COMPUTER ENGINEERING (ICMCCE 2020), 2020, : 50 - 56