Search papers, labs, and topics across Lattice.
This paper introduces SuperFlex, an advanced framework for point cloud decomposition that enhances the traditional superquadrics by incorporating bending and tapering deformations. By implementing a novel loss formulation, SuperFlex achieves significant improvements in reconstruction accuracy and robustness, particularly when dealing with partial point clouds. Experimental results show that SuperFlex outperforms existing optimization- and learning-based methods while retaining a compact representation of 3D objects.
SuperFlex achieves unprecedented reconstruction accuracy for 3D point clouds by enabling deformable superquadrics to represent complex geometries robustly.
Superquadrics have proven to provide a compact, geometrically meaningful representation for 3D objects. However, existing methods suffer from limited reconstruction accuracy, are restricted to rigid primitives, and lack robustness to partial point clouds. In this work, we present SuperFlex, an enhanced framework that expands the expressive power and applicability of superquadric decompositions. First, we introduce a novel loss formulation which significantly improves reconstruction accuracy. Second, we include bending and tapering deformations, enabling high-fidelity representation of curved and asymmetric geometries. Finally, we leverage these high-quality decompositions as supervision to train a model that is robust to partial real-world point clouds. Experiments demonstrate substantial improvements in reconstruction accuracy over both optimization- and learning-based baselines while maintaining a highly compact primitive representation.
ETH
Stanford HAI