3D Point Capsule Networks

Zhao, Yanyun; Birdal, Tolga; Deng, Haowen; Tombari, Federico

doi:10.1109/cvpr.2019.00110

Cited by 330 publications

(223 citation statements)

References 35 publications

Supporting

Mentioning

222

Contrasting

Order By: Relevance

“…AtlasNet [26] extends the FoldingNet to multiple grid patches whereas SO-Net [39] aggregates the point features into SOM node features to encode the spatial distributions. PointCapsNet [96] introduces an autoencoder based on dynamic routing to extract latent capsules and a few MLPs that generate multiple point patches from the latent capsules with distinct grids.…”

Section: Deep Learning On Point Cloudsmentioning

confidence: 99%

“…We also report part classification accuracy. Following [96], we randomly sample 1% and 5% of the ShapeNetPart train set to evaluate the point features in a semi-supervised setting. We use the same pre-trained model to extract the point features of the sampled training data, along with validation and test samples without any finetuning.…”

Section: Part Segmentationmentioning

confidence: 99%

See 1 more Smart Citation

Unsupervised Multi-Task Feature Learning on Point Clouds

Hassani

Haley

2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

176

117

View full text Add to dashboard Cite

We introduce an unsupervised multi-task model to jointly learn point and shape features on point clouds. We define three unsupervised tasks including clustering, reconstruction, and self-supervised classification to train a multi-scale graph-based encoder. We evaluate our model on shape classification and segmentation benchmarks. The results suggest that it outperforms prior state-of-the-art unsupervised models: In the ModelNet40 classification task, it achieves an accuracy of 89.1% and in ShapeNet segmentation task, it achieves an mIoU of 68.2 and accuracy of 88.6%.

show abstract

Section: Deep Learning On Point Cloudsmentioning

confidence: 99%

Section: Part Segmentationmentioning

confidence: 99%

Unsupervised Multi-Task Feature Learning on Point Clouds

Hassani

Haley

2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

176

117

View full text Add to dashboard Cite

show abstract

“…Our evaluation is mainly against supervised methods since there has been hardly any semi-supervised segmentation methods that take only a few exemplars and segment shapes in a whole set. An exception is the very recent work by Zhao et al [69] which used only 5% of the training data. In comparison, their IOU is 70% averaged over the shapes, while our 1-exemplar result is 73.5% even by setting all IOUs of cars to zero.…”

Section: One-shot Training Vs Supervised Methodsmentioning

confidence: 99%

BAE-NET: Branched Autoencoder for Shape Co-Segmentation

Chen

Yin

Fisher

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

123

112

View full text Add to dashboard Cite

We treat shape co-segmentation as a representation learning problem and introduce BAE-NET, a branched autoencoder network, for the task. The unsupervised BAE-NET is trained with a collection of un-segmented shapes, using a shape reconstruction loss, without any ground-truth labels. Specifically, the network takes an input shape and encodes it using a convolutional neural network, whereas the decoder concatenates the resulting feature code with a point coordinate and outputs a value indicating whether the point is inside/outside the shape. Importantly, the decoder is branched: each branch learns a compact representation for one commonly recurring part of the shape collection, e.g., airplane wings. By complementing the shape reconstruction loss with a label loss, BAE-NET is easily tuned for one-shot learning. We show unsupervised, weakly supervised, and one-shot learning results by BAE-NET, demonstrating that using only a couple of exemplars, our network can generally outperform state-of-the-art supervised methods trained on hundreds of segmented shapes. Code is available at https://github.com/czq142857/BAE-NET.

show abstract

“…Recent trends in data driven approaches have encouraged the researchers to harness deep learning to surmount these nuisances. Representative works include 3DMatch [56], PPFNet [20], CGF [33], 3D-FeatNet [29], PPF-FoldNet [19] and 3D point-capsule networks [57], all outperforming the handcrafted alternatives by large margin. While the descriptors in 2D are typically complemented by the useful information of local orientation, derived from the local image appearance [37], the nature of 3D data renders the task of finding a unique and consistent local coordinate frame way more challenging [24,42].…”

Section: Related Workmentioning

confidence: 99%

3D Local Features for Direct Pairwise Registration

Deng¹,

Birdal

Ilić³

2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

Self Cite

115

View full text Add to dashboard Cite

We present a novel, data driven approach for solving the problem of registration of two point cloud scans. Our approach is direct in the sense that a single pair of corresponding local patches already provides the necessary transformation cue for the global registration. To achieve that, we first endow the state of the art PPF-FoldNet [19] auto-encoder (AE) with a pose-variant sibling, where the discrepancy between the two leads to pose-specific descriptors. Based upon this, we introduce RelativeNet, a relative pose estimation network to assign correspondence-specific orientations to the keypoints, eliminating any local reference frame computations. Finally, we devise a simple yet effective hypothesize-and-verify algorithm to quickly use the predictions and align two point sets. Our extensive quantitative and qualitative experiments suggests that our approach outperforms the state of the art in challenging real datasets of pairwise registration and that augmenting the keypoints with local pose information leads to better generalization and a dramatic speed-up.

show abstract

3D Point Capsule Networks

Cited by 330 publications

References 35 publications

Unsupervised Multi-Task Feature Learning on Point Clouds

Unsupervised Multi-Task Feature Learning on Point Clouds

BAE-NET: Branched Autoencoder for Shape Co-Segmentation

3D Local Features for Direct Pairwise Registration

Contact Info

Product

Resources

About