GeoNet: Deep Geodesic Networks for Point Cloud Analysis
This work addresses the challenge of analyzing point clouds for computer vision and graphics applications by incorporating surface topology, offering incremental improvements over existing methods.
The authors tackled the problem of modeling surface connectivity in point clouds, which is lost in raw data, by introducing GeoNet, a deep learning architecture that learns geodesic-aware representations. Their method improved state-of-the-art results on tasks like point upsampling, normal estimation, mesh reconstruction, and non-rigid shape classification.
Surface-based geodesic topology provides strong cues for object semantic analysis and geometric modeling. However, such connectivity information is lost in point clouds. Thus we introduce GeoNet, the first deep learning architecture trained to model the intrinsic structure of surfaces represented as point clouds. To demonstrate the applicability of learned geodesic-aware representations, we propose fusion schemes which use GeoNet in conjunction with other baseline or backbone networks, such as PU-Net and PointNet++, for down-stream point cloud analysis. Our method improves the state-of-the-art on multiple representative tasks that can benefit from understandings of the underlying surface topology, including point upsampling, normal estimation, mesh reconstruction and non-rigid shape classification.