Point2Skeleton: Learning Skeletal Representations from Point Clouds

Abstract

We introduce Point2Skeleton, an unsupervised method to learn skeletal representations from point clouds. Existing skeletonization methods are limited to tubular shapes and the stringent requirement of watertight input, while our method aims to produce more generalized skeletal representations for complex structures and handle point clouds. Our key idea is to use the insights of the medial axis transform (MAT) to capture the intrinsic geometric and topological natures of the original input points. We first predict a set of skeletal points by learning a geometric transformation, and then analyze the connectivity of the skeletal points to form skeletal mesh structures. Extensive evaluations and comparisons show our method has superior performance and robustness. The learned skeletal representation will benefit several unsupervised tasks for point clouds, such as surface reconstruction and segmentation.

Walk in the Cloud: Learning Curves for Point Clouds Shape Analysis

“…To reach a better understanding of the 3D targets, effective point cloud analysis techniques and methods are in great demand. With the thriving of deep learning, the pioneer works [26,28] and their followers [20,6,43,1,42,7,17,47] processed point clouds through well-designed neural networks to learn the latent mappings between input point coordinates and the ground truth labels. Differing from conventional 2D vision tasks, the points are usually in irregular and unordered forms, hence, the effective designs of feature aggregation and message passing schemes among point clouds still remains challenging.…”