Explaining the Ambiguity of Object Detection and 6D Pose From Visual Data

@article{Manhardt2018ExplainingTA,
  title={Explaining the Ambiguity of Object Detection and 6D Pose From Visual Data},
  author={Fabian Manhardt and Diego Mart{\'i}n Arroyo and C. Rupprecht and Benjamin Busam and Nassir Navab and Federico Tombari},
  journal={2019 IEEE/CVF International Conference on Computer Vision (ICCV)},
  year={2018},
  pages={6840-6849}
}
3D object detection and pose estimation from a single image are two inherently ambiguous problems. Oftentimes, objects appear similar from different viewpoints due to shape symmetries, occlusion and repetitive textures. This ambiguity in both detection and pose estimation means that an object instance can be perfectly described by several different poses and even classes. In this work we propose to explicitly deal with these ambiguities. For each object instance we predict multiple 6D pose… 
View on IEEE
arxiv.org
77 Citations
Highly Influential Citations
5
Background Citations
53
Methods Citations
25
Results Citations
1

77 Citations

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

This work devise a network to reconstruct the three rotation axis primitive images of a target object and predict the underlying uncertainty along each primitive axis and optimize multi-object poses using visual measurements and camera poses by treating it as an object SLAM problem.

On Object Symmetries and 6D Pose Estimation from Images

It is explained why symmetrical objects can be a challenge when training machine learning algorithms that aim at estimating their 6D pose from images, and an efficient and simple solution that relies on the normalization of the pose rotation is proposed.

Ki-Pode: Keypoint-based Implicit Pose Distribution Estimation of Rigid Objects

This work proposes a novel pose distribution estimation method based on an implicit formulation of the probability distribution over object pose derived from an intermediary representation of an object as a set of keypoints, which leads to reliable estimates.

EPOS: Estimating 6D Pose of Objects With Symmetries

A new method for estimating the 6D pose of rigid objects with available 3D models from a single RGB input image using a robust and efficient variant of the PnP-RANSAC algorithm, which outperforms all RGB and most RGB-D and D methods on the T-LESS and LM-O datasets.

Learning Implicit Probability Distribution Functions for Symmetric Orientation Estimation from RGB Images Without Pose Labels

This work designs a two-stage pipeline consisting of point cloud registration and render-and-compare validation to generate multiple symmetrical pseudo-ground-truth pose labels for each image, and introduces an automatic pose labeling scheme.

Robust 6D Object Pose Estimation by Learning RGB-D Features

This work proposes a novel discrete- continuous formulation for rotation regression to resolve this local-optimum problem, and shows that the proposed method outperforms state-of-the-art approaches.

CPS: Class-level 6D Pose and Shape Estimation From Monocular Images

This paper proposes the first deep learning approach for class-wise monocular 6D pose estimation, coupled with metric shape retrieval, and proposes a new loss formulation which directly optimizes over all parameters, i.e. 3D orientation, translation, scale and shape at the same time.

Pose Ambiguity Elimination Algorithm for 3C Components Assembly Pose Estimation in Point Cloud

A pose ambiguity elimination algorithm based on PCA (Principal Component Analysis) and 2D image template matching is proposed and has higher accuracy than traditional methods and the efficiency meets the need.

Learning Orientation Distributions for Object Pose Estimation

This work proposes two learned methods for estimating a distribution over an object’s orientation that give the best performance on objects with unknown symmetries, accurately modeling both symmetric and non-symmetric objects, without any requirement of symmetry annotation.

DSC-PoseNet: Learning 6DoF Object Pose Estimation via Dual-scale Consistency

This paper presents a two-step pose estimation framework to attain 6DoF object poses from 2D object bounding-boxes and designs a dual-scale pose estimation network, namely DSC-PoseNet, to predict object poses by employing a differential renderer.
...

References

SHOWING 1-10 OF 65 REFERENCES

On Evaluation of 6D Object Pose Estimation

The paper defines 6D object pose estimation problems, proposes an evaluation methodology and introduces three new pose error functions that deal with pose ambiguity, compared with functions commonly used in the literature and shown to remove certain types of non-intuitive outcomes.

Learning 6D Object Pose Estimation Using 3D Object Coordinates

This work addresses the problem of estimating the 6D Pose of specific objects from a single RGB-D image by presenting a learned, intermediate representation in form of a dense 3D object coordinate labelling paired with a dense class labelling.

Uncertainty-Driven 6D Pose Estimation of Objects and Scenes from a Single RGB Image

A regularized, auto-context regression framework is developed which iteratively reduces uncertainty in object coordinate and object label predictions and an efficient way to marginalize object coordinate distributions over depth is introduced to deal with missing depth information.

Detection and fine 3D pose estimation of texture-less objects in RGB-D images

Experimental evaluation shows that the proposed method yields a recognition rate comparable to the state of the art, while its complexity is sub-linear in the number of templates.

DeepIM: Deep Iterative Matching for 6D Pose Estimation

A novel deep neural network for 6D pose matching named DeepIM is proposed, trained to predict a relative pose transformation using a disentangled representation of 3D location and 3D orientation and an iterative training process.

PoseCNN: A Convolutional Neural Network for 6D Object Pose Estimation in Cluttered Scenes

This work introduces PoseCNN, a new Convolutional Neural Network for 6D object pose estimation, which is highly robust to occlusions, can handle symmetric objects, and provide accurate pose estimation using only color images as input.

BB8: A Scalable, Accurate, Robust to Partial Occlusion Method for Predicting the 3D Poses of Challenging Objects without Using Depth

A novel method for 3D object detection and pose estimation from color images only that uses segmentation to detect the objects of interest in 2D even in presence of partial occlusions and cluttered background and is the first to report results on the Occlusion dataset using color imagesonly.

Hashmod: A Hashing Method for Scalable 3D Object Detection

It is shown empirically that the complexity of the method is sublinear with the number of objects and it enables detection and pose estimation of many 3D objects with high accuracy while outperforming the state-of-the-art in terms of runtime.

Latent-Class Hough Forests for 3D Object Detection and Pose Estimation

A novel framework, Latent-Class Hough Forests, for 3D object detection and pose estimation in heavily cluttered and occluded scenes using the state-of-the-art template matching feature, LINEMOD, into a scale-invariant patch descriptor and integrated into a regression forest using a novel template-based split function.

Fast Detection of Multiple Textureless 3-D Objects

A fast edge-based approach for detection and approximate pose estimation of multiple textureless objects in a single image that outperforms the state-of-the-art approach by Damen et al. (2012).
...