GEOMetrics: Exploiting Geometric Structure for Graph-Encoded Objects

Abstract

Mesh models are a promising approach for encoding the structure of 3D objects. Current mesh reconstruction systems predict uniformly distributed vertex locations of a predetermined graph through a series of graph convolutions, leading to compromises with respect to performance or resolution. In this paper, we argue that the graph representation of geometric objects allows for additional structure, which should be leveraged for enhanced reconstruction. Thus, we propose a system which properly benefits from the advantages of the geometric structure of graph encoded objects by introducing (1) a graph convolutional update preserving vertex information; (2) an adaptive splitting heuristic allowing detail to emerge; and (3) a training objective operating both on the local surfaces defined by vertices as well as the global structure defined by the mesh. Our proposed method is evaluated on the task of 3D object reconstruction from images with the ShapeNet dataset, where we demonstrate state of the art performance, both visually and numerically, while having far smaller space requirements by generating adaptive meshes

18 pages

Cited by in corpus (16)

• Learning to Predict 3D Objects with an Interpolation-based Differentiable Renderer
• Differentiable Rendering: A Survey
• Accelerating 3D Deep Learning with PyTorch3D
• Joint Hand-object 3D Reconstruction from a Single Image with Cross-branch Feature Fusion
• Kaolin: A PyTorch Library for Accelerating 3D Deep Learning Research
• 3D Shape Reconstruction from Vision and Touch
• Higher-Order Function Networks for Learning Composable 3D Object Representations
• Local Deep Implicit Functions for 3D Shape
• Neural Mesh Flow: 3D Manifold Mesh Generation via Diffeomorphic Flows
• Self-Sampling for Neural Point Cloud Consolidation
• Holistic 3D Scene Understanding from a Single Image with Implicit Representation
• Active 3D Shape Reconstruction from Vision and Touch
• STD-Net: Structure-preserving and Topology-adaptive Deformation Network for 3D Reconstruction from a Single Image
• Shelf-Supervised Mesh Prediction in the Wild
• Absolute 3D Pose Estimation and Length Measurement of Severely Deformed Fish from Monocular Videos in Longline Fishing
• GAMesh: Guided and Augmented Meshing for Deep Point Networks