Recognizing human actions from low-resolution videos by region-based mixture models.
Ying ZhaoHuijun DiJian ZhangYao LuFeng LvPublished in: ICME (2016)
Keyphrases
- mixture model
- low resolution
- recognizing human actions
- action recognition
- high resolution
- super resolution
- human actions
- em algorithm
- gaussian mixture model
- generative model
- probabilistic model
- expectation maximization
- model selection
- maximum likelihood
- high quality
- language model
- image patches
- image segmentation
- unsupervised learning
- spatio temporal
- face images
- computer vision
- image processing
- probability density function
- activity recognition
- video sequences
- test images
- space time
- learning algorithm
- data streams