Pose-guided multi-task video transformer for driver action recognition.
Ricardo PizarroRoberto ValleLuis Miguel BergasaJosé Miguel BuenaposadaLuis BaumelaPublished in: CoRR (2024)
Keyphrases
- action recognition
- multi task
- action detection
- human actions
- action classification
- human pose
- recognizing actions
- multi task learning
- video dataset
- recognizing human actions
- static images
- learning tasks
- human activities
- recognition of human actions
- activity recognition
- space time interest points
- bag of words
- computer vision
- human detection
- multi class
- feature selection
- learning problems
- transfer learning
- pose estimation
- gaussian processes
- body parts
- motion history images
- video sequences
- max margin
- multimedia
- human motion
- video frames
- atomic actions
- learning algorithm