Neural Object Learning for 6D Pose Estimation Using a Few Cluttered Images

Park, Kiru; Patten, Timothy; Vincze, Markus

doi:10.1007/978-3-030-58548-8_38

Cited by 15 publications

(10 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Labeled w/ GT 6D Pose is a tedious, costly and labor-intensive chore. Another category of works [25,34,41] makes use of the available object model to reduce the labeling effort. They render images with the CAD model and train the pose estimator on the synthetic images.…”

Section: Training Inferencementioning

confidence: 99%

Weakly Supervised Learning of Keypoints for 6D Object Pose Estimation

Tian¹,

Lee²

2022

Preprint

View full text Add to dashboard Cite

State-of-the-art approaches for 6D object pose estimation require large amounts of labeled data to train the deep networks. However, the acquisition of 6D object pose annotations is tedious and labor-intensive in large quantity. To alleviate this problem, we propose a weakly supervised 6D object pose estimation approach based on 2D keypoint detection. Our method trains only on image pairs with known relative transformations between their viewpoints. Specifically, we assign a set of arbitrarily chosen 3D keypoints to represent each unknown target 3D object and learn a network to detect their 2D projections that comply with the relative camera viewpoints. During inference, our network first infers the 2D keypoints from the query image and a given labeled reference image. We then use these 2D keypoints and the arbitrarily chosen 3D keypoints retained from training to infer the 6D object pose. Extensive experiments demonstrate that our approach achieves comparable performance with state-of-the-art fully supervised approaches.

show abstract

Section: Training Inferencementioning

confidence: 99%

Weakly Supervised Learning of Keypoints for 6D Object Pose Estimation

Tian¹,

Lee²

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Each predicted hypothesis corresponded to a single 3D translation and rotation, and estimated hypotheses collapsed onto the same valid pose when the object appearance was unique. Other recent variations of this method include Trabelsi et al [45], who used a multi-task CNN-based encoder/multi-decoder network, and Wang et al [48] and [20,34,43], who used a rendering method by a self-supervised model on unannotated real RGB-D data to find a visually and geometrically optimal alignment.…”

Section: Related Workmentioning

confidence: 99%

“…There were five different network architectures proposed in the intial ResNet paper [13]. While some 6 DoF pose recovery works use variations of ResNet-18 [36,47,50,45] others use ResNet-50 [48,34]. Some customize the structure by converting it to encoder [45,34,50,47], adding extra layers and skip connections [36] while others use the original ResNet unaltered [14,33].…”

Section: D1 Number Of Skip Connectionsmentioning

confidence: 99%

Vote from the Center: 6 DoF Pose Estimation in RGB-D Images by Radial Keypoint Voting

Zand

Etemad

et al. 2021

Preprint

View full text Add to dashboard Cite

We propose a novel keypoint voting scheme based on intersecting spheres, that is more accurate than existing schemes and allows for a smaller set of more disperse keypoints. The scheme forms the basis of the proposed RCV-Pose method for 6 DoF pose estimation of 3D objects in RGB-D data, which is particularly effective at handling occlusions. A CNN is trained to estimate the distance between the 3D point corresponding to the depth mode of each RGB pixel, and a set of 3 disperse keypoints defined in the object frame. At inference, a sphere of radius equal to this estimated distance is generated, centered at each 3D point. The surface of these spheres votes to increment a 3D accumulator space, the peaks of which indicate keypoint locations. The proposed radial voting scheme is more accurate than previous vector or offset schemes, and robust to disperse keypoints. Experiments demonstrate RCVPose to be highly accurate and competitive, achieving state-of-the-art results on LINEMOD (99.7%), YCB-Video (97.2%) datasets, and notably scoring +7.9% higher than previous methods on the challenging Occlusion LINEMOD (71.1%) dataset.

show abstract

“…Without depth information, these methods suffer dramatic performance degradation. NOL [28] generates novel views of a given object from a few cluttered real-world images in which object poses are needed. The generated views are self-supervised by the input images.…”

Section: Related Workmentioning

confidence: 99%

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

Yang¹,

Yu²,

Yang³

2021

Preprint

View full text Add to dashboard Cite

Compared to 2D object bounding-box labeling, it is very difficult for humans to annotate 3D object poses, especially when depth images of scenes are unavailable. This paper investigates whether we can estimate the object poses effectively when only RGB images and 2D object annotations are given. To this end, we present a two-step pose estimation framework to attain 6DoF object poses from 2D object bounding-boxes. In the first step, the framework learns to segment objects from real and synthetic data in a weaklysupervised fashion, and the segmentation masks will act as a prior for pose estimation. In the second step, we design a dual-scale pose estimation network, namely DSC-PoseNet, to predict object poses by employing a differential renderer. To be specific, our DSC-PoseNet firstly predicts object poses in the original image scale by comparing the segmentation masks and the rendered visible object masks. Then, we resize object regions to a fixed scale to estimate poses once again. In this fashion, we eliminate large scale variations and focus on rotation estimation, thus facilitating pose estimation. Moreover, we exploit the initial pose estimation to generate pseudo ground-truth to train our DSC-PoseNet in a self-supervised manner. The estimation results in these two scales are ensembled as our final pose estimation. Extensive experiments on widely-used benchmarks demonstrate that our method outperforms state-of-the-art models trained on synthetic data by a large margin and even is on par with several fully-supervised methods.

show abstract

Neural Object Learning for 6D Pose Estimation Using a Few Cluttered Images

Cited by 15 publications

References 44 publications

Weakly Supervised Learning of Keypoints for 6D Object Pose Estimation

Weakly Supervised Learning of Keypoints for 6D Object Pose Estimation

Vote from the Center: 6 DoF Pose Estimation in RGB-D Images by Radial Keypoint Voting

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

Contact Info

Product

Resources

About