The shape variational autoencoder: A deep generative model of part-segmented 3D objects.
Charlie NashChristopher K. I. WilliamsPublished in: Comput. Graph. Forum (2017)
Keyphrases
- generative model
- d objects
- shape descriptors
- cad models
- image contours
- d mesh
- curved objects
- object parts
- probabilistic model
- geometric models
- pose estimation
- recognition and pose estimation
- dirichlet process mixture models
- multi view
- object recognition
- topological information
- three dimensional
- viewpoint
- range data
- prior knowledge
- primal sketch
- em algorithm
- shape analysis
- object features
- conditional random fields
- spin images
- line drawings
- shape matching
- bayesian framework
- expectation maximization
- topic models
- shape model
- object categories
- cad model
- maximum likelihood
- semi supervised
- multiscale