TearingNet: Point Cloud Autoencoder to Learn Topology-Friendly Representations
This work addresses the problem of topology representation in point clouds for researchers in geometric deep learning, offering a novel method but with incremental improvements over existing approaches.
The paper tackles the challenge of capturing complex topologies in point cloud data by proposing TearingNet, an autoencoder that learns explicit topology through iterative tearing and folding modules, resulting in faithful reconstructions and superior performance compared to benchmarks.
Topology matters. Despite the recent success of point cloud processing with geometric deep learning, it remains arduous to capture the complex topologies of point cloud data with a learning model. Given a point cloud dataset containing objects with various genera, or scenes with multiple objects, we propose an autoencoder, TearingNet, which tackles the challenging task of representing the point clouds using a fixed-length descriptor. Unlike existing works directly deforming predefined primitives of genus zero (e.g., a 2D square patch) to an object-level point cloud, our TearingNet is characterized by a proposed Tearing network module and a Folding network module interacting with each other iteratively. Particularly, the Tearing network module learns the point cloud topology explicitly. By breaking the edges of a primitive graph, it tears the graph into patches or with holes to emulate the topology of a target point cloud, leading to faithful reconstructions. Experimentation shows the superiority of our proposal in terms of reconstructing point clouds as well as generating more topology-friendly representations than benchmarks.