3D-PRNN: Generating Shape Primitives with Recurrent Neural Networks
arXiv:1708.01648
Abstract
The success of various applications including robotics, digital content creation, and visualization demand a structured and abstract representation of the 3D world from limited sensor data. Inspired by the nature of human perception of 3D shapes as a collection of simple parts, we explore such an abstract shape representation based on primitives. Given a single depth image of an object, we present 3D-PRNN, a generative recurrent neural network that synthesizes multiple plausible shapes composed of a set of primitives. Our generative model encodes symmetry characteristics of common man-made objects, preserves long-range structural coherence, and describes objects of varying complexity with a compact representation. We also propose a method based on Gaussian Fields to generate a large scale dataset of primitive-based shape representations to train our network. We evaluate our approach on a wide range of examples and show that it outperforms nearest-neighbor based shape retrieval methods and is on-par with voxel-based generative models while using a significantly reduced parameter space.
ICCV 2017
Cited by in corpus (6)
- Supervised Fitting of Geometric Primitives to 3D Point Clouds
- Blood Vessel Geometry Synthesis using Generative Adversarial Networks
- Learning Manifold Patch-Based Representations of Man-Made Shapes
- RNN-based Generative Model for Fine-Grained Sketching
- Physical Primitive Decomposition
- Generic Primitive Detection in Point Clouds Using Novel Minimal Quadric Fits