PrimA6D: Rotational Primitive Reconstruction for Enhanced and Robust 6D Pose Estimation

Jeon, Myung–Hwan; Kim, Ayoung

doi:10.1109/lra.2020.3004322

Cited by 10 publications

(8 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2) Rotational Primitive Axis Reconstruction: Two major challenges in object pose estimation are occlusion and symmetry. To overcome these challenges, we leverage ideas from our previous work [10] on primitive reconstruction for objects followed by keypoint extraction from these primitives. By doing so, we can reliably estimate the pose from even small and texture-less objects.…”

Section: D Object Pose Initializationmentioning

confidence: 99%

“…Similarly, as in [10], we cropp the input RGB image and concentrate on the target object using the segmentation map. We then transform the cropped RGB image into the primitive image using the Auto Encoder (AE)-based primitive reconstruction network.…”

Section: D Object Pose Initializationmentioning

confidence: 99%

“…Unlike in previous studies, we introduce an object ambiguity-aware object pose estimation network, PrimA6D++, as a generic uncertainty estimation. Our previous work, PrimA6D [10], introduced the rotation primitive reconstruction which predicts color-coded orientation axes for robust object pose estimation, yet this method cannot handle the object ambiguity and uncertainty of the predicted pose. Extending from [10], the proposed network predicts three rotation axis The cropped RGB image using the object segmentation mask transforms into the three rotation axis primitive images.…”

Section: Introductionmentioning

confidence: 99%

“…Our previous work, PrimA6D [10], introduced the rotation primitive reconstruction which predicts color-coded orientation axes for robust object pose estimation, yet this method cannot handle the object ambiguity and uncertainty of the predicted pose. Extending from [10], the proposed network predicts three rotation axis The cropped RGB image using the object segmentation mask transforms into the three rotation axis primitive images. In each rotation axis primitive image, 14 keypoints and their associated uncertainties are estimated.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Ambiguity-Aware Multi-Object Pose Optimization for Visually-Assisted Robot Manipulation

Jeon

Kim

Ryu

et al. 2023

IEEE Robot. Autom. Lett.

Self Cite

View full text Add to dashboard Cite

6D object pose estimation aims to infer the relative pose between the object and the camera using a single image or multiple images. Most works have focused on predicting the object pose without associated uncertainty under occlusion and structural ambiguity (symmetricity). However, these works demand prior information about shape attributes, and this condition is hardly satisfied in reality; even asymmetric objects may be symmetric under the viewpoint change. In addition, acquiring and fusing diverse sensor data is challenging when extending them to robotics applications. Tackling these limitations, we present an ambiguity-aware 6D object pose estimation network, PrimA6D++, as a generic uncertainty prediction method. The major challenges in pose estimation, such as occlusion and symmetry, can be handled in a generic manner based on the measured ambiguity of the prediction. Specifically, we devise a network to reconstruct the three rotation axis primitive images of a target object and predict the underlying uncertainty along each primitive axis. Leveraging the estimated uncertainty, we then optimize multi-object poses using visual measurements and camera poses by treating it as an object SLAM problem. The proposed method shows a significant performance improvement in T-LESS and YCB-Video datasets. We further demonstrate real-time scene recognition capability for visually-assisted robot manipulation. Our code and supplementary materials are available at https://github.com/rpmsnu/PrimA6D.

show abstract

Section: D Object Pose Initializationmentioning

confidence: 99%

Section: D Object Pose Initializationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Ambiguity-Aware Multi-Object Pose Optimization for Visually-Assisted Robot Manipulation

Jeon

Kim

Ryu

et al. 2023

IEEE Robot. Autom. Lett.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Finally, our approach to tactile pose estimation is related to methods recently explored in the computer vision community where they render realistic images of objects and learn how to estimate the orientation of an object given a new image of it [33,34]. While in vision, the most likely estimate is often sufficient, in tactile sensing, different object poses are more likely to produce the same observation.…”

Section: Related Workmentioning

confidence: 99%

Tactile Object Pose Estimation from the First Touch with Geometric Contact Rendering

Bauza,

Valls,

Lim

et al. 2020

Preprint

View full text Add to dashboard Cite

In this paper, we present an approach to tactile pose estimation from the first touch for known objects. First, we create an object-agnostic map from real tactile observations to contact shapes. Next, for a new object with known geometry, we learn a tailored perception model completely in simulation. To do so, we simulate the contact shapes that a dense set of object poses would produce on the sensor. Then, given a new contact shape obtained from the sensor output, we match it against the pre-computed set using the object-specific embedding learned purely in simulation using contrastive learning. This results in a perception model that can localize objects from a single tactile observation. It also allows reasoning over pose distributions and including additional pose constraints coming from other perception systems or multiple contacts. We provide quantitative results for four objects. Our approach provides high accuracy pose estimations from distinctive tactile observations while regressing pose distributions to account for those contact shapes that could result from different object poses. We further extend and test our approach in multi-contact scenarios where several tactile sensors are simultaneously in contact with the object. Website: mcube.mit.edu/research/tactile_loc_first_touch.html

show abstract

A Lightweight Model for Satellite Pose Estimation

Carcagnì¹,

Leo²,

Spagnolo³

et al. 2022

Image Analysis and Processing – ICIAP 2022

View full text Add to dashboard Cite

PrimA6D: Rotational Primitive Reconstruction for Enhanced and Robust 6D Pose Estimation

Cited by 10 publications

References 22 publications

Ambiguity-Aware Multi-Object Pose Optimization for Visually-Assisted Robot Manipulation

Ambiguity-Aware Multi-Object Pose Optimization for Visually-Assisted Robot Manipulation

Tactile Object Pose Estimation from the First Touch with Geometric Contact Rendering

A Lightweight Model for Satellite Pose Estimation

Contact Info

Product

Resources

About