CSMMI: Class-Specific Maximization of Mutual Information for Action and Gesture Recognition.
Jun WanVassilis AthitsosPat JangyodsukHugo Jair EscalanteQiuqi RuanIsabelle GuyonPublished in: IEEE Trans. Image Process. (2014)
Keyphrases
- gesture recognition
- class specific
- maximization of mutual information
- image super resolution
- hidden markov models
- human computer interaction
- image classification
- object class
- image registration
- dictionary learning
- object categories
- image processing
- object detection
- object classes
- multiple features
- superpixels
- sparse representation
- image representation
- object recognition