· 2018
FoldingNet: Point Cloud Auto-Encoder via Deep Grid Deformation
- 1,362citations
- 2018year
Short summary
FoldingNet, a novel auto-encoder for point clouds, uses a graph-enhanced encoder and a folding decoder that deforms a 2D grid onto 3D object surfaces, achieving low reconstruction errors with 7% of parameters compared to fully-connected networks.
AI-generated from the title and abstract; the full text is not read.
Key points
- Introduces FoldingNet, an auto-encoder for unsupervised learning on point clouds.
- Encoder incorporates graph-based enhancement for local structure promotion.
- Decoder deforms a 2D grid onto 3D object surfaces for reconstruction.
- Achieves low reconstruction errors with ~7% of parameters compared to fully-connected decoders.
- Leads to more discriminative representations and higher linear SVM classification accuracy.
AI-generated from the title and abstract; the full text is not read.
Abstract
Recent deep networks that directly handle points in a point set, e.g., PointNet, have been state-of-the-art for supervised learning tasks on point clouds such as classification and segmentation. In this work, a novel end-to-end deep auto-encoder is proposed to address unsupervised learning challenges on point clouds. On the encoder side, a graph-based enhancement is enforced to promote local structures on top of PointNet. Then, a novel folding-based decoder deforms a canonical 2D grid onto the underlying 3D object surface of a point cloud, achieving low reconstruction errors even for objects with delicate structures. The proposed decoder only uses about 7% parameters of a decoder with fully-connected neural networks, yet leads to a more discriminative representation that achieves higher linear SVM classification accuracy than the benchmark. In addition, the proposed decoder structure is shown, in theory, to be a generic architecture that is able to reconstruct an arbitrary point cloud from a 2D grid. Our code is available at http://www.merl.com/research/license#FoldingNet.
The authors' abstract, as published at the source., 2018 · DOI ↗
Continue with a free account
Ask the paper: 3 free questions a day about this paper; save it, get its citation, new summaries every day for your field. Takeaways are Premium.
Continue free on the webSign in with Google or Apple; no card needed. You come back to this paper.
On your phone:
Field: Computational Mechanics
Computational MechanicsEngineering