Human Activity Recognition Based on Motion Projection Profile Features in Surveillance Videos Using Support Vector Machines and Gaussian Mixture Models.
J. ArunnehruM. Kalaiselvi GeethaPublished in: SSCC (2015)
Keyphrases
- human activity recognition
- gaussian mixture model
- surveillance videos
- human activities
- activity recognition
- feature vectors
- mixture model
- feature space
- video surveillance
- optical flow
- low level
- human computer interaction
- motion detection
- video data
- image sequences
- probabilistic model
- multi camera
- action recognition
- expectation maximization
- motion estimation
- key frames
- d objects
- object recognition
- training data