Pedestrian Path, Pose and Intention Prediction through Gaussian Process Dynamical Models and Pedestrian Activity Recognition.
Raúl QuinteroIgnacio ParraDavid Fernández LlorcaMiguel Ángel SoteloPublished in: CoRR (2020)
Keyphrases
- gaussian process
- activity recognition
- dynamical models
- dynamical model
- pose tracking
- gaussian processes
- regression model
- human activities
- action recognition
- bayesian framework
- hyperparameters
- latent variables
- model selection
- dynamical systems
- semi supervised
- visual tracking
- latent space
- incremental learning
- human motion
- sample size
- human pose estimation
- mean shift
- d objects
- supervised learning
- subject specific
- bayesian inference
- video sequences
- motion model