Unsupervised Feature Learning of Human Actions As Trajectories in Pose Embedding Manifold.
Jogendra Nath KunduMaharshi GorPhani Krishna UppalaVenkatesh Babu RadhakrishnanPublished in: WACV (2019)
Keyphrases
- human actions
- unsupervised feature learning
- body pose
- action recognition
- human pose
- recognizing actions
- view invariant
- deep learning
- spatio temporal
- higher order
- space time
- human activities
- human motion
- pose estimation
- activity recognition
- latent space
- video sequences
- motion patterns
- human body
- visual features
- body parts
- appearance model
- moving objects
- manifold learning
- machine learning
- motion trajectories
- three dimensional
- joint angles
- computer vision
- image sequences
- face recognition
- expectation maximization
- image classification
- d objects
- feature space