A multi-feature selection approach for gender identification of handwriting based on kernel mutual information.
Ning BiChing Y. SuenNicola NobileJun TanPublished in: Pattern Recognit. Lett. (2019)
Keyphrases
- mutual information
- feature selection
- conditional mutual information
- information theoretic
- support vector
- feature space
- parzen window
- information gain
- image registration
- similarity measure
- class separability
- text categorization
- feature extraction
- kernel methods
- multiple kernel learning
- statistical dependence
- text classification
- character recognition
- medical image registration
- normalized mutual information
- writer identification
- information theoretic measures
- machine learning
- feature selection algorithms
- model selection
- feature set
- support vector machine
- selection criterion
- individual differences
- handwriting recognition
- selected features
- dimensionality reduction
- face recognition
- data sets
- multimodal image registration
- mutual information maximization