Joint object recognition and pose estimation using a nonlinear view-invariant latent generative model.
Amr BakryTarek El-GaalyMohamed ElhoseinyAhmed M. ElgammalPublished in: WACV (2016)
Keyphrases
- generative model
- view invariant
- object recognition and pose estimation
- probabilistic model
- action recognition
- multi view
- pose estimation
- single view
- latent space
- bayesian framework
- latent structure
- generative process
- latent variables
- em algorithm
- human actions
- human motion
- discriminative learning
- prior knowledge
- expectation maximization
- viewpoint
- spatio temporal
- latent dirichlet allocation
- semi supervised
- topic models
- discriminative models
- three dimensional
- motion capture
- hidden markov models