Human action recognition based on GMM-UBM supervector using SVM with non-linear GMM KL and GUMI.
Ngoc Nam BuiYoung J. KimPublished in: ICDIP (2015)
Keyphrases
- gaussian mixture model
- action recognition
- feature vectors
- mixture model
- static images
- human movements
- human actions
- human activities
- activity recognition
- em algorithm
- human detection
- spatial temporal
- bag of words
- support vector
- maximum likelihood
- deformable part models
- recognizing human actions
- max margin
- action classification
- feature space
- body parts
- computer vision
- background subtraction
- support vector machine svm
- motion history images
- expectation maximization
- human object interactions
- human pose
- view invariant
- video dataset
- motion capture data
- recognizing actions
- depth sensors
- action recognition in videos
- multi class
- support vector machine
- knn
- bag of features
- model selection
- recognition of human actions
- feature extraction