Training a classifier for activity recognition using body motion simulation.
Michelangelo GrossoDavide LenaSalvatore RinaudoDavid Alejandro Fernandez GuzmanDanilo DemarchiPublished in: BioCAS (2017)
Keyphrases
- semi supervised
- training process
- supervised learning
- classifier training
- training examples
- training samples
- active learning
- training set
- svm classifier
- motion estimation
- human activity recognition
- human motion
- training data
- class labels
- motion capture
- image sequences
- supervised training
- optical flow
- training dataset
- human body
- body movements
- simulation model
- feature space
- data sets
- motion patterns
- activity analysis
- motion field
- motion analysis
- motion model
- learning algorithm
- camera motion
- decision trees
- classification performances
- boosted classifiers
- support vector
- sufficient training data
- feature selection and classifier
- multi layer perceptron
- humanoid robot
- motion segmentation
- classification method
- space time
- feature points
- object detection
- feature selection
- neural network