Neural varifolds: an aggregate representation for quantifying geometry of point clouds

Juheon Lee, Xiaohao Cai, Carola-Bibiane Schönlieb, Simon Masnou

Primary Area: metric learning, kernel learning, and sparse coding

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Keywords: surface geometry, point clouds, mesh, neural tangent kernel, varifold

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.

Abstract: Point clouds are popular 3D representations for real-life objects (such as in LiDAR, Kinect and smartphones) due to their detailed and compact representation of surface-based geometry. Recent approaches characterise the geometry of point clouds by bringing deep-learning based techniques together with geometric fidelity metrics such as optimal transportation costs (e.g. Chamfer and Wasserstein metrics). In this paper, we propose a new surface geometry characterisation within this realm, namely a neural varifold representation of point clouds. Here the surface is represented as a measure/distribution over both point positions and tangent spaces of point clouds. The varifold representation not only helps to quantify the surface geometry of point clouds through the manifold-based discrimination, but also subtle geometric consistency on the surface due to the combined product space. This study proposes neural varifold algorithms to compute varifold norm between two point clouds using neural networks on point clouds and their neural tangent kernel representations. The proposed neural varifold is evaluated on three different tasks -- shape classification, shape reconstruction and shape matching. Detailed evaluation and comparison to the state-of-the-art methods demonstrate that the proposed versatile neural varifold is superior in shape classification particularly for limited data and is quite competitive for shape reconstruction and matching.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 3927