Filters








551 Hits in 6.6 sec

Shape Prior Deformation for Categorical 6D Object Pose and Size Estimation [article]

Meng Tian, Marcelo H Ang Jr, Gim Hee Lee
2020 arXiv   pre-print
Additionally, our network infers the dense correspondences between the depth observation of the object instance and the reconstructed 3D model to jointly estimate the 6D object pose and size.  ...  We design an autoencoder that trains on a collection of object models and compute the mean latent embedding for each category to learn the categorical shape priors.  ...  This research is supported in parts by the Singapore MOE Tier 1 grant R-252-000-A65-114, and the Agency for Science, Technology and Research (A*STAR) under its AME Programmatic Funding Scheme (Project  ... 
arXiv:2007.08454v1 fatcat:efjmfukibjhezc45sccwvpafqy

Category-Level 6D Object Pose and Size Estimation using Self-Supervised Deep Prior Deformation Networks [article]

Jiehong Lin, Zewei Wei, Changxing Ding, Kui Jia
2022 arXiv   pre-print
DPDN learns to deform features of categorical shape priors to match those of object observations, and is thus able to establish deep correspondence in the feature space for direct regression of object  ...  It is difficult to precisely annotate object instances and their semantics in 3D space, and as such, synthetic data are extensively used for these tasks, e.g., category-level 6D object pose and size estimation  ...  DPDN deforms categorical shape priors in the feature space to pair with object observations, and establishes deep correspondence for direct estimates of object poses and sizes; upon DPDN, a novel self-supervised  ... 
arXiv:2207.05444v2 fatcat:z5ycm4bd6jasvndoguo3efxs44

Category-Level 6D Object Pose Estimation in the Wild: A Semi-Supervised Learning Approach and A New Dataset [article]

Yang Fu, Xiaolong Wang
2022 arXiv   pre-print
6D object pose estimation is one of the fundamental problems in computer vision and robotics research.  ...  We utilize this data to generalize category-level 6D object pose estimation in the wild with semi-supervised learning.  ...  The RePoNet is composed of two branches of networks with a Pose Network to estimate the 6D object pose and a Shape Network to estimate the 3D object shape.  ... 
arXiv:2206.15436v1 fatcat:qygwkazj7vetbpa25qj4mwx3te

6D-ViT: Category-Level 6D Object Pose Estimation via Transformer-based Instance Representation Learning [article]

Lu Zou, Zhangjin Huang, Naijie Gu, Guoping Wang
2021 arXiv   pre-print
This paper presents 6D-ViT, a transformer-based instance representation learning network, which is suitable for highly accurate category-level object pose estimation on RGB-D images.  ...  Finally, the instance 6D pose is computed by leveraging the correspondence among dense representations, shape priors, and the instance point clouds.  ...  matrix and deformation field) used to calculate the 6D object pose.  ... 
arXiv:2110.04792v2 fatcat:xyyldbzpgzayhaqxsdt4wycrg4

FS-Net: Fast Shape-based Network for Category-Level 6D Object Pose Estimation with Decoupled Rotation Mechanism [article]

Wei Chen, Xi Jia, Hyung Jin Chang, Jinming Duan, Linlin Shen, Ales Leonardis
2021 arXiv   pre-print
To tackle this problem, we propose a fast shape-based network (FS-Net) with efficient category-level feature extraction for 6D pose estimation.  ...  In this paper, we focus on category-level 6D pose and size estimation from monocular RGB-D image.  ...  Shape-Prior [35] estimated the object size and 6D pose from dense-fusion feature [40] , while we estimate the pose from point cloud feature.  ... 
arXiv:2103.07054v2 fatcat:fxdjf2xiq5dm5jiwweuxaebxcu

Category-Level 6D Object Pose Estimation via Cascaded Relation and Recurrent Reconstruction Networks [article]

Jiaze Wang, Kai Chen, Qi Dou
2021 arXiv   pre-print
shape prior.  ...  Precisely recovering instance 3D model in the canonical space and accurately matching it with the observation is an essential point when estimating 6D pose for unseen objects.  ...  The work was supported by the Hong Kong Centre for Logistics Robotics.  ... 
arXiv:2108.08755v1 fatcat:svsczdz2fng67kn5nc6gjhtxfe

FS-Net: Fast Shape-based Network for Category-Level 6D Object Pose Estimation with Decoupled Rotation Mechanism

Wei Chen, Xi Jia, Hyung Jin Chang, Jinming Duan, Linlin Shen, Ales Leonardis
2021 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
In this paper, we focus on category-level 6D pose and size estimation from a monocular RGB-D image.  ...  For translation and size, we estimate them by two residuals: the difference between the mean of object points and ground truth translation, and the difference between the mean size of the category and  ...  The authors also thank Tze Ho Elden Tse for his valuable comments.  ... 
doi:10.1109/cvpr46437.2021.00163 fatcat:gbqdoli2znefzolfakgg2vh4za

CenterSnap: Single-Shot Multi-Object 3D Shape Reconstruction and Categorical 6D Pose and Size Estimation [article]

Muhammad Zubair Irshad, Thomas Kollar, Michael Laskey, Kevin Stone, Zsolt Kira
2022 arXiv   pre-print
Through extensive experiments, we demonstrate that our approach significantly outperforms all shape completion and categorical 6D pose and size estimation baselines on multi-object ShapeNet and NOCS datasets  ...  Hence, we present a simple one-stage approach to predict both the 3D shape and estimate the 6D pose and size jointly in a bounding-box free manner.  ...  Datasets: We utilize the NOCS [22] dataset to evaluate both shape reconstruction and categorical 6D pose and size estimation.  ... 
arXiv:2203.01929v1 fatcat:si5et7k5wnaupkfkut4jdogmk4

TransGrasp: Grasp Pose Estimation of a Category of Objects by Transferring Grasps from Only One Labeled Instance [article]

Hongtao Wen, Jianhang Yan, Wanli Peng, Yi Sun
2022 arXiv   pre-print
Specifically, we perform grasp pose transfer across a category of objects based on their shape correspondences and propose a grasp pose refinement module to further fine-tune grasp pose of grippers so  ...  Grasp pose estimation is an important issue for robots to interact with the real world.  ...  This research was supported by the National Natural Science Foundation of China (No.U1708263 and No.61873046).  ... 
arXiv:2207.07861v3 fatcat:pl6sjgujd5b6tbtad36lqruddm

SDFEst: Categorical Pose and Shape Estimation of Objects from RGB-D using Signed Distance Fields [article]

Leonard Bruns, Patric Jensfelt
2022 arXiv   pre-print
In this paper, we present a modular pipeline for pose and shape estimation of objects from RGB-D images given their category.  ...  The core of our method is a generative shape model, which we integrate with a novel initialization network and a differentiable renderer to enable 6D pose and shape estimation from a single or multiple  ...  The authors thank Raghav Bongole for his contributions to the software repository.  ... 
arXiv:2207.04880v1 fatcat:y6ou7vcahbck7pug4inb6tcsgu

On the Evaluation of RGB-D-based Categorical Pose and Shape Estimation [article]

Leonard Bruns, Patric Jensfelt
2022 arXiv   pre-print
Recently, various methods for 6D pose and shape estimation of objects have been proposed.  ...  Typically, these methods evaluate their pose estimation in terms of average precision, and reconstruction quality with chamfer distance.  ...  Furthermore, this evaluation procedure can be adjusted for categorical pose estimation by using d and δ only, and for categorical pose and size estimation by using IoU instead of F ∆ .  ... 
arXiv:2202.10346v1 fatcat:vcn7ixgy3zgodosou5cigaskm4

CATRE: Iterative Point Clouds Alignment for Category-level Object Pose Refinement [article]

Xingyu Liu, Gu Wang, Yi Li, Xiangyang Ji
2022 arXiv   pre-print
object shape and color, etc.  ...  Given an initial pose estimate, CATRE predicts a relative transformation between the initial pose and ground truth by means of aligning the partially observed point cloud and an abstract shape prior.  ...  Notably, SPD [47] proposes to extract a categorical shape prior and adapt it to various instances via deformation prediction, in an effort to improve the matching of correspondences.  ... 
arXiv:2207.08082v1 fatcat:zrrzscz5bvcohjiziq7vvnmxca

Accurate 6D Object Pose Estimation by Pose Conditioned Mesh Reconstruction [article]

Pedro Castro, Anil Armagan, Tae-Kyun Kim
2019 arXiv   pre-print
Current 6D object pose methods consist of deep CNN models fully optimized for a single object but with its architecture standardized among objects with different shapes.  ...  In contrast to previous works, we explicitly exploit each object's distinct topological information i.e. 3D dense meshes in the pose estimation model, with an automated process and prior to any post-processing  ...  This work was in part financially supported via student consultation for Remark Holdings, Inc.  ... 
arXiv:1910.10653v1 fatcat:r4lxtpjecjbbvcdzqg5jyj2snu

A Survey on Joint Object Detection and Pose Estimation using Monocular Vision [article]

Aniruddha V Patil, Pankaj Rabha
2018 arXiv   pre-print
In this survey we present a complete landscape of joint object detection and pose estimation methods that use monocular vision.  ...  These descriptors or models include chordiograms, shape-aware deformable parts model, bag of boundaries, distance transform templates, natural 3D markers and facet features whereas the estimation methods  ...  Prior knowledge of the facets of the object of interest is then used for detection and pose estimation [1] .  ... 
arXiv:1811.10216v1 fatcat:4rs7zhi4jffapneubmopjcjtpe

Pixel2Mesh++: 3D Mesh Generation and Refinement from Multi-View Images [article]

Chao Wen, Yinda Zhang, Chenjie Cao, Zhuwen Li, Xiangyang Xue, Yanwei Fu
2022 arXiv   pre-print
Model analysis experiments show that our model is robust to the quality of the initial mesh and the error of camera pose, and can be combined with a differentiable renderer for test-time optimization.  ...  While many previous works learn to hallucinate the shape directly from priors, we adopt to further improve the shape quality by leveraging cross-view information with a graph convolution network.  ...  [69] show that the discontinuity of the rotation representation is the main reason for the difficulty in pose estimation, and proposed a continuous 6D rotation representation. Xu et al.  ... 
arXiv:2204.09866v1 fatcat:y6uea56dvfgw3fwuus4ckilnxa
« Previous Showing results 1 — 15 out of 551 results