MX-LSTM: Mixing Tracklets and Vislets to Jointly Forecast Trajectories and Head Poses.
Irtiza HasanFrancesco SettiTheodore TsesmelisAlessio Del BueFabio GalassoMarco CristaniPublished in: CVPR (2018)
Keyphrases
- motion patterns
- multi target tracking
- object trajectories
- recurrent neural networks
- spatio temporal
- data association
- multiple targets
- crowded scenes
- short term
- real time
- trajectory data
- space time
- motion model
- appearance model
- motion analysis
- phase space
- image sequences
- neural network
- humanoid robot
- multiple cameras
- computer vision
- bayesian framework
- long term
- three dimensional