Learning Shape Abstractions by Assembling Volumetric Primitives
Authors:
- Shubham TulsianiSTShubham Tulsiani
- Hao SuHSHao Su
- Leonidas J. GuibasLGLeonidas J. Guibas
- Alexei A. EfrosAEAlexei A. Efros
- Jitendra MalikJMJitendra Malik
- 1 July 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 10636919,p. 1466-1474
Abstract
We present a learning framework for abstracting complex shapes by learning to assemble objects using 3D volumetric primitives. In addition to generating simple and geometrically interpretable explanations of 3D objects, our framework also allows us to automatically discover and exploit consistent structure in the data. We demonstrate that using our method allows predicting shape representations which can be leveraged for obtaining a consistent parsing across the instances of a shape collection and constructing an interpretable shape similarity measure. We also examine applications for image-based prediction as well as shape manipulation.Keywords
All Related Versions
This publication has 23 references indexed in Scilit:
- Data-driven structural priors for shape completionACM Transactions on Graphics, 2015
- Single-view reconstruction via joint analysis of image and shape collectionsACM Transactions on Graphics, 2015
- Meta-representation of shape familiesACM Transactions on Graphics, 2014
- Recurring part arrangements in shape collectionsComputer Graphics Forum, 2014
- Active co-analysis of a set of shapesACM Transactions on Graphics, 2012
- A probabilistic model for component-based shape synthesisACM Transactions on Graphics, 2012
- Unsupervised co-segmentation of a set of shapes via descriptor-space spectral clustering2011
- Blocks World Revisited: Image Understanding Using Qualitative Geometry and Mechanics2010
- Simple statistical gradient-following algorithms for connectionist reinforcement learningMachine Learning, 1992
- Recognition-by-components: A theory of human image understanding.Psychological Review, 1987