Action Recognition From Video Using Feature Covariance Matrices.
Kai GuoPrakash IshwarJanusz KonradPublished in: IEEE Trans. Image Process. (2013)
Keyphrases
- action recognition
- covariance matrices
- motion features
- spatio temporal interest points
- human actions
- action classification
- covariance matrix
- video dataset
- action detection
- activity recognition
- feature vectors
- recognizing human actions
- maximum likelihood
- human activities
- recognition of human actions
- bag of words
- distance measure
- space time interest points
- computer vision
- vector space
- gaussian mixture model
- body parts
- gaussian distribution
- video data
- motion history images
- global features
- video sequences
- image features
- gaussian mixture
- spatio temporal
- bayesian networks
- image processing
- video frames
- query processing
- principal component analysis
- atomic actions
- dimensionality reduction
- feature selection
- learning algorithm