EMAG: An Extended Multimodal Attribute Grammar for Behavioural Features

This work aims at providing a linguistic representation of multimodal behavioural features in human-human and human-machine communication processes. Different multimodal behavioural features are considered, such as the tone and type of sentence for speech; size, slant, and individual social stile for handwriting; and trait, pressure, dimension, and individual personality for sketch. In order to represent these features, an Extended Multimodal Attribute Grammar (EMAG) has been defined in the article. A discussion of the EMAG expressiveness and a validation of the EMAG inference have been provided.