TY - GEN
T1 - PointTriNet: Learned Triangulation of 3D Point Sets
AU - Sharp, Nicholas
AU - Ovsjanikov, Maks
N1 - KAUST Repository Item: Exported on 2022-06-30
Acknowledged KAUST grant number(s): CRG-2017-3426
Acknowledgements: The authors are grateful to Marie-Julie Rakotosaona and Keenan Crane for fruitful initial discussions, and to Angela Dai for assistance comparing with Scan2Mesh. Parts of this work were supported by an NSF Graduate Research Fellowship, the KAUST OSR Award No. CRG-2017-3426 and the ERC Starting Grant No. 758800 (EXPROTEA).
This publication acknowledges KAUST support, but has no KAUST affiliated authors.
PY - 2020/11/3
Y1 - 2020/11/3
N2 - This work considers a new task in geometric deep learning: generating a triangulation among a set of points in 3D space. We present PointTriNet, a differentiable and scalable approach enabling point set triangulation as a layer in 3D learning pipelines. The method iteratively applies two neural networks: a classification network predicts whether a candidate triangle should appear in the triangulation, while a proposal network suggests additional candidates. Both networks are structured as PointNets over nearby points and triangles, using a novel triangle-relative input encoding. Since these learning problems operate on local geometric data, our method is efficient and scalable, and generalizes to unseen shape categories. Our networks are trained in an unsupervised manner from a collection of shapes represented as point clouds. We demonstrate the effectiveness of this approach for classical meshing tasks, robustness to outliers, and as a component in end-to-end learning systems.
AB - This work considers a new task in geometric deep learning: generating a triangulation among a set of points in 3D space. We present PointTriNet, a differentiable and scalable approach enabling point set triangulation as a layer in 3D learning pipelines. The method iteratively applies two neural networks: a classification network predicts whether a candidate triangle should appear in the triangulation, while a proposal network suggests additional candidates. Both networks are structured as PointNets over nearby points and triangles, using a novel triangle-relative input encoding. Since these learning problems operate on local geometric data, our method is efficient and scalable, and generalizes to unseen shape categories. Our networks are trained in an unsupervised manner from a collection of shapes represented as point clouds. We demonstrate the effectiveness of this approach for classical meshing tasks, robustness to outliers, and as a component in end-to-end learning systems.
UR - http://hdl.handle.net/10754/679514
UR - https://link.springer.com/10.1007/978-3-030-58592-1_45
UR - http://www.scopus.com/inward/record.url?scp=85097388821&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-58592-1_45
DO - 10.1007/978-3-030-58592-1_45
M3 - Conference contribution
SN - 9783030585914
SP - 762
EP - 778
BT - Computer Vision – ECCV 2020
PB - Springer International Publishing
ER -