Ginevra Castellano 




                                                                                                                    Home | Research | Projects | Activities | Publications | Contact                                                                                                                     

Research

Affect sensitive human-robot interaction

Robots are increasingly being studied as partners that collaborate and do things with people. Societal needs and new technological opportunities will make robots, in the near future, more and more available to large amounts of users directly in homes, offices, schools and hospitals environments. This will require robots to be endowed with social skills, for example the ability to engage users via new means of natural interaction, such as gestures, motion and other forms of behavioural expressions, and to adapt to their feelings and emotions. I am interested in the exploration of how the perception of users’ affective and social behaviour can help robots engage with users and behave in a socially acceptable way over extended periods of time.
 
Robot
Automatic analysis of non-verbal behaviour for multimodal affect recognition

The recent rise of behavioural computing has brought a lot of attention to the design of systems for the automatic analysis of human behaviour in computer-based applications. My research addresses the automatic analysis of multimodal human behaviour (e.g., facial expressions, eye gaze, body movements and postures) and contextual information for affect recognition. I am particularly interested in exploring the role of social and task context and how they can be modelled for the purpose of automatically predicting affect in human-computer and human-robot interaction.

Face tracking

Affective interaction based on body movement


My research in this areas addresses the role of body movements and their quality as a channel to recognise and communicate emotions in human-computer and human-robot interaction. This research starts from the assumption that it is possible to discriminate different affective, emotional states not only from the type of body movement or gesture performed, but also from the way it is performed.


Affective loop